----------------------------------------------------------------------------------------------------------
       log:  C:\Temp\DS2000replicationlog.log
  log type:  text
 opened on:  22 Apr 2006, 11:09:55

. 
. do C:\Temp\ISQ_dofile.txt

. 
. 
. **  Do-file for "No Easy Choices: Estimating the Effects of United Nations Peacekeeping (Response to Kin
> g and Zeng)"
. **  Nicholas Sambanis and Michael W. Doyle
. **  April 17, 2006
. 
. 
. ** This file contains the do-files to replicate all analyses done in STATA version 8
. ** The code and results for analyses done in GAUSS will be presented separately
. 
. 
. 
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. 
. * Table 1 in the supplement
. * Regressions of Peacebuilding Outcomes from DS2000
. 
. * Regression 1 (Model A, Table 2 in DS2000)
. logistic pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap un2cint treaty develop exp decade, clus
> ter(cluster)

Logistic regression                               Number of obs   =        122
                                                  Wald chi2(11)   =      30.42
                                                  Prob > chi2     =     0.0014
Log pseudo-likelihood = -45.236556                Pseudo R2       =     0.4286

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   .1245494   .0824476    -3.15   0.002      .034031    .4558361
     logcost |   .5859226   .0866553    -3.61   0.000     .4384816    .7829412
      wardur |   1.009152   .0056258     1.63   0.102     .9981859    1.020239
     factnum |   .1713298   .1175107    -2.57   0.010       .04467    .6571283
    factnum2 |   1.121169   .0634538     2.02   0.043     1.003451    1.252696
    trnsfcap |   1.003813   .0015836     2.41   0.016     1.000714    1.006922
     un2cint |     2.0275   .5504967     2.60   0.009     1.190822    3.452033
      treaty |   8.017213   6.119424     2.73   0.006     1.796029    35.78767
     develop |   1.000577   .0003618     1.60   0.111     .9998683    1.001287
         exp |    .004806   .0143451    -1.79   0.074     .0000138    1.669022
      decade |   .7053552   .1233589    -2.00   0.046      .500659    .9937421
------------------------------------------------------------------------------

. 
. 
. * Regression 2 (Model A4, Table 3 in DS2000)
. logistic pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap un2int treaty develop exp decade, clust
> er(cluster)

Logistic regression                               Number of obs   =        122
                                                  Wald chi2(11)   =      31.13
                                                  Prob > chi2     =     0.0010
Log pseudo-likelihood = -46.473584                Pseudo R2       =     0.4130

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   .1332634   .0872723    -3.08   0.002     .0369208    .4810064
     logcost |   .6180468   .0832412    -3.57   0.000     .4746549    .8047571
      wardur |   1.008356   .0053468     1.57   0.117     .9979306     1.01889
     factnum |   .2095493   .1354812    -2.42   0.016     .0590139    .7440782
    factnum2 |   1.106277   .0608399     1.84   0.066     .9932349    1.232185
    trnsfcap |   1.003701   .0015405     2.41   0.016     1.000686    1.006724
      un2int |   3.823865   2.413391     2.13   0.034     1.109879    13.17435
      treaty |   8.940362   6.722024     2.91   0.004     2.048137    39.02575
     develop |   1.000738   .0003938     1.87   0.061     .9999663     1.00151
         exp |   .0031959    .009516    -1.93   0.054     9.33e-06    1.094237
      decade |   .6771509   .1274593    -2.07   0.038      .468236    .9792782
------------------------------------------------------------------------------

. 
. 
. * Regression 3 (Model A8, Table 3 from DS2000)
. logistic pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap untype4 treaty develop exp decade, clus
> ter(cluster)

Logistic regression                               Number of obs   =        122
                                                  Wald chi2(11)   =      41.47
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.648672                Pseudo R2       =     0.4234

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |    .175169   .1066165    -2.86   0.004     .0531342    .5774845
     logcost |   .6409204   .0804494    -3.54   0.000     .5011409    .8196876
      wardur |   1.006406   .0056767     1.13   0.258     .9953414    1.017594
     factnum |   .2839568    .199722    -1.79   0.073     .0715412    1.127062
    factnum2 |   1.063611   .0695978     0.94   0.346     .9355868    1.209154
    trnsfcap |   1.004102   .0015905     2.58   0.010     1.000989    1.007224
     untype4 |   22.99287   25.08811     2.87   0.004     2.709104    195.1465
      treaty |   8.367608   6.871941     2.59   0.010       1.6732     41.8461
     develop |    1.00079   .0004278     1.85   0.065     .9999523    1.001629
         exp |    .002439   .0074913    -1.96   0.050     5.93e-06    1.003837
      decade |   .7417257   .1254109    -1.77   0.077     .5325032    1.033152
------------------------------------------------------------------------------

. 
. 
. * In note 6 in the ISQ piece, we mention that our model A from DS2000 uses a variable that codes the typ
> e of UN mandate (un2cint).  That variable is coded as follows: 
. 
. tab un2cint

      Type of UN |
        Presence |      Freq.     Percent        Cum.
-----------------+-----------------------------------
            None |         77       63.11       63.11
       Mediation |          9        7.38       70.49
        Observer |         17       13.93       84.43
             PKO |         15       12.30       96.72
     Enforcement |          4        3.28      100.00
-----------------+-----------------------------------
           Total |        122      100.00

. 
. * There is also a second version, which excludes mediations (not usually thought of as peacekeeping miss
> ions)
. 
. tab untype

  Type of UN |
ops -- break |
   PKO & PBO |      Freq.     Percent        Cum.
-------------+-----------------------------------
        None |         86       70.49       70.49
    Observer |         17       13.93       84.43
     TradPKO |          8        6.56       90.98
    MultiPKO |          7        5.74       96.72
Enforcement+ |          4        3.28      100.00
-------------+-----------------------------------
       Total |        122      100.00

. 
. 
. 
. 
. ** Table 2 in the supplement: Peacebuilding Models With and Without Interaction Terms
. ** Two regressions: Model with interactions first and then original model
. ** Likelihood ratio test follows (dropping clustering)
. 
. * interaction between war duration and MultiPKO has already been created (gen un4wdur=untype4*wardur)
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap treaty develop exp decade untype4 un4wdur,
>  nolog

Logit estimates                                   Number of obs   =        122
                                                  LR chi2(12)     =      68.54
                                                  Prob > chi2     =     0.0000
Log likelihood = -44.902028                       Pseudo R2       =     0.4329

------------------------------------------------------------------------------
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.666386   .5962555    -2.79   0.005    -2.835025   -.4977468
     logcost |  -.4371467   .1361314    -3.21   0.001    -.7039594   -.1703341
      wardur |   .0055304   .0041052     1.35   0.178    -.0025156    .0135765
     factnum |  -1.045376   1.112292    -0.94   0.347    -3.225428    1.134676
    factnum2 |   .0322828   .1454715     0.22   0.824    -.2528361    .3174016
    trnsfcap |   .0040839   .0021462     1.90   0.057    -.0001225    .0082903
      treaty |   2.126205   .7726536     2.75   0.006     .6118316    3.640578
     develop |   .0007643   .0004609     1.66   0.097    -.0001392    .0016677
         exp |  -6.215339   3.559492    -1.75   0.081    -13.19181    .7611367
      decade |   -.283677   .1916737    -1.48   0.139    -.6593505    .0919965
     untype4 |   .2616384   2.820808     0.09   0.926    -5.267043     5.79032
     un4wdur |   .0372672   .0498987     0.75   0.455    -.0605324    .1350667
       _cons |   8.189294    2.46752     3.32   0.001     3.353044    13.02554
------------------------------------------------------------------------------

. 
. lrtest, saving(0)

. 
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap treaty develop exp decade untype4 ,  nolog

Logit estimates                                   Number of obs   =        122
                                                  LR chi2(11)     =      67.05
                                                  Prob > chi2     =     0.0000
Log likelihood = -45.648672                       Pseudo R2       =     0.4234

------------------------------------------------------------------------------
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.742004   .5966751    -2.92   0.004    -2.911466   -.5725426
     logcost |  -.4448499   .1366485    -3.26   0.001    -.7126762   -.1770237
      wardur |   .0063857   .0039946     1.60   0.110    -.0014434    .0142149
     factnum |  -1.258933   .8257796    -1.52   0.127    -2.877432    .3595649
    factnum2 |   .0616696   .0930059     0.66   0.507    -.1206187    .2439579
    trnsfcap |   .0040934   .0020585     1.99   0.047     .0000589     .008128
      treaty |   2.124368   .7785935     2.73   0.006      .598353    3.650383
     develop |   .0007901   .0004655     1.70   0.090    -.0001224    .0017025
         exp |  -6.016179   3.552358    -1.69   0.090    -12.97867    .9463152
      decade |  -.2987758   .1925219    -1.55   0.121    -.6761118    .0785601
     untype4 |   3.135184   1.493477     2.10   0.036     .2080233    6.062345
       _cons |   8.608862   2.196549     3.92   0.000     4.303705    12.91402
------------------------------------------------------------------------------

. 
. lrtest
You ran lrtest using the old syntax.  Click here to learn about the new syntax.


likelihood-ratio test                                  LR chi2(1)  =      1.49
(Assumption: . nested in LRTEST_0)                     Prob > chi2 =    0.2217

. 
. 
. 
. ** Comparisons of fit statistics between the original and modified models:
. 
. * Original Model:
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 , cluste
> r(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -50.186571
Iteration 2:   log pseudo-likelihood = -46.266465
Iteration 3:   log pseudo-likelihood = -45.668048
Iteration 4:   log pseudo-likelihood = -45.648703
Iteration 5:   log pseudo-likelihood = -45.648672

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(11)   =      41.47
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.648672                Pseudo R2       =     0.4234

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.742004   .6086493    -2.86   0.004    -2.934935   -.5490736
     logcost |  -.4448499   .1255217    -3.54   0.000    -.6908679    -.198832
      wardur |   .0063857   .0056405     1.13   0.258    -.0046695     .017441
     factnum |  -1.258933   .7033537    -1.79   0.073    -2.637481    .1196146
    factnum2 |   .0616696   .0654354     0.94   0.346    -.0665814    .1899205
    trnsfcap |   .0040934    .001584     2.58   0.010     .0009888    .0071981
     develop |   .0007901   .0004274     1.85   0.065    -.0000477    .0016278
         exp |  -6.016179    3.07149    -1.96   0.050    -12.03619      .00383
      decade |  -.2987758   .1690799    -1.77   0.077    -.6301664    .0326148
      treaty |   2.124368   .8212551     2.59   0.010     .5147377    3.733999
     untype4 |   3.135184   1.091125     2.87   0.004     .9966179    5.273751
       _cons |   8.608862     2.1571     3.99   0.000     4.381024     12.8367
------------------------------------------------------------------------------

. 
. fitstat, saving(fit1)

Measures of Fit for logit of pbs2s3

Log-Lik Intercept Only:      -79.173     Log-Lik Full Model:          -45.649
D(110):                       91.297     LR(11):                       67.048
                                         Prob > LR:                     0.000
McFadden's R2:                 0.423     McFadden's Adj R2:             0.272
Maximum Likelihood R2:         0.423     Cragg & Uhler's R2:            0.582
McKelvey and Zavoina's R2:     0.671     Efron's R2:                    0.501
Variance of y*:                9.989     Variance of error:             3.290
Count R2:                      0.844     Adj Count R2:                  0.558
AIC:                           0.945     AIC*n:                       115.297
BIC:                        -437.145     BIC':                        -14.204

(Indices saved in matrix fs_fit1)

. 
. 
. * Modified Model:
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4wdur,
>  cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -49.666147
Iteration 2:   log pseudo-likelihood = -45.718462
Iteration 3:   log pseudo-likelihood = -44.995036
Iteration 4:   log pseudo-likelihood = -44.914883
Iteration 5:   log pseudo-likelihood = -44.903002
Iteration 6:   log pseudo-likelihood = -44.902039
Iteration 7:   log pseudo-likelihood = -44.902028
Iteration 8:   log pseudo-likelihood = -44.902028

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      53.21
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.902028                Pseudo R2       =     0.4329

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.666386   .6058667    -2.75   0.006    -2.853863   -.4789092
     logcost |  -.4371467    .125279    -3.49   0.000    -.6826891   -.1916044
      wardur |   .0055304   .0058193     0.95   0.342    -.0058752     .016936
     factnum |  -1.045376   .8994968    -1.16   0.245    -2.808357    .7176056
    factnum2 |   .0322828   .1040289     0.31   0.756    -.1716101    .2361756
    trnsfcap |   .0040839   .0017085     2.39   0.017     .0007353    .0074325
     develop |   .0007643   .0004192     1.82   0.068    -.0000574     .001586
         exp |  -6.215339   3.065318    -2.03   0.043    -12.22325   -.2074256
      decade |   -.283677   .1687572    -1.68   0.093    -.6144351    .0470811
      treaty |   2.126205   .8017543     2.65   0.008     .5547954    3.697614
     untype4 |   .2616384   1.392347     0.19   0.851    -2.467311    2.990588
     un4wdur |   .0372672   .0110637     3.37   0.001     .0155826    .0589517
       _cons |   8.189294    2.35083     3.48   0.000     3.581751    12.79684
------------------------------------------------------------------------------

. 
. fitstat, using(fit1)

Measures of Fit for logit of pbs2s3

                             Current            Saved       Difference
Model:                         logit            logit
N:                               122              122                0
Log-Lik Intercept Only:      -79.173          -79.173            0.000
Log-Lik Full Model:          -44.902          -45.649            0.747
D:                            89.804(109)      91.297(110)      -1.493(-1)
LR:                           68.541(12)       67.048(11)        1.493(1)
Prob > LR:                     0.000            0.000            0.000
McFadden's R2:                 0.433            0.423            0.009
McFadden's Adj R2:             0.269            0.272           -0.003
Maximum Likelihood R2:         0.430            0.423            0.007
Cragg & Uhler's R2:            0.591            0.582            0.010
McKelvey and Zavoina's R2:     0.722            0.671            0.051
Efron's R2:                    0.504            0.501            0.003
Variance of y*:               11.831            9.989            1.842
Variance of error:             3.290            3.290            0.000
Count R2:                      0.836            0.844           -0.008
Adj Count R2:                  0.535            0.558           -0.023
AIC:                           0.949            0.945            0.004
AIC*n:                       115.804          115.297            0.507
BIC:                        -433.834         -437.145            3.311
BIC':                        -10.893          -14.204            3.311

Difference of    3.311 in BIC' provides positive support for saved model.

. 
. * results should show Difference of 3.311 in BIC' -- positive support for saved (original) model
. 
. ** see supplement posted online for results using different permutations of the model (with more interac
> tion terms)
. 
. 
. 
. ** The ISQ piece mentions that the LR test drops clustering. See discussion of this in the supplement
. ** We also mention that MultiPKO still significant with bootstrapped standard errors, dropping clusterin
> g by country
. ** Results obtained as follows:
. 
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 1234

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4" "_b
> [wartype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _b
> [treaty] _b[untype4] ", reps(1000) saving(C:\Temp\boot_multipko) replace 

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unty
> pe4
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[untype4]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -1.742004 -.1774103  1.025133  -3.753666   .2696575   (N)
             |                                       -4.38067  -.3641894   (P)
             |                                      -4.182204  -.2026127  (BC)
       _bs_2 |  1000 -.4448499 -.0935132  .2351202  -.9062361   .0165363   (N)
             |                                      -1.089128  -.2058999   (P)
             |                                      -.8028194  -.1253831  (BC)
       _bs_3 |  1000  .0063857  .0015042  .0078979  -.0091126    .021884   (N)
             |                                      -.0040312   .0272999   (P)
             |                                      -.0045348   .0261055  (BC)
       _bs_4 |  1000 -1.258933  .0195987  2.122546  -5.424094   2.906228   (N)
             |                                      -4.402498   2.368955   (P)
             |                                      -4.031449   2.709319  (BC)
       _bs_5 |  1000  .0616696 -.0410322  .3368514  -.5993479    .722687   (N)
             |                                      -.5097139   .4244343   (P)
             |                                      -.4765692   .4390881  (BC)
       _bs_6 |  1000  .0040934  .0022541  .0064662  -.0085954   .0167823   (N)
             |                                      -.0002533   .0242621   (P)
             |                                      -.0018702   .0115202  (BC)
       _bs_7 |  1000  .0007901  .0002201  .0006798  -.0005439    .002124   (N)
             |                                      -.0000177   .0025867   (P)
             |                                       -.000255   .0019786  (BC)
       _bs_8 |  1000 -6.016179 -1.488614  4.844132  -15.52202   3.489661   (N)
             |                                      -19.04841    .634542   (P)
             |                                       -14.8587   2.680867  (BC)
       _bs_9 |  1000 -.2987758 -.0230262  .2507745  -.7908809   .1933293   (N)
             |                                      -.8648311   .1114595   (P)
             |                                      -.8138463   .1314881  (BC)
      _bs_10 |  1000  2.124368  .3686049   1.31426  -.4546594   4.703396   (N)
             |                                        .280693   5.162487   (P)
             |                                      -.7439417   4.356194  (BC)
      _bs_11 |   649  3.135184  .2413619  1.569352   .0535544   6.216814   (N)
             |                                       1.058917    7.26387   (P)
             |                                       1.072016   7.531678  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 1234

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn" "_
> b[wartype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _
> b[treaty] _b[unintrvn] ", reps(1000) saving(C:\Temp\boot_unintrvn) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unin
> trvn
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[unintrvn]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.036098  -.310538  1.170562   -4.33314   .2609437   (N)
             |                                      -4.968345  -.7008833   (P)
             |                                        -4.1829  -.3876035  (BC)
       _bs_2 |  1000 -.5545111 -.1165735  .3025237  -1.148166   .0391437   (N)
             |                                      -1.401597  -.2740033   (P)
             |                                      -1.078451  -.2191278  (BC)
       _bs_3 |  1000  .0087491   .001869  .0081701  -.0072834   .0247817   (N)
             |                                      -.0008658   .0309268   (P)
             |                                      -.0013795   .0295728  (BC)
       _bs_4 |  1000 -1.648769 -.5958874  1.918034  -5.412606   2.115068   (N)
             |                                      -6.161785  -.0466028   (P)
             |                                      -3.962131   1.643141  (BC)
       _bs_5 |  1000  .1056245  .0499806  .2535151  -.3918586   .6031077   (N)
             |                                      -.0787192   .5650981   (P)
             |                                       -.389236   .3717726  (BC)
       _bs_6 |  1000  .0037728  .0016351  .0056654  -.0073446   .0148902   (N)
             |                                      -.0007647   .0212454   (P)
             |                                      -.0025072   .0118241  (BC)
       _bs_7 |  1000  .0005694  .0001944  .0006254  -.0006579   .0017967   (N)
             |                                      -.0001565   .0020883   (P)
             |                                      -.0003142   .0017858  (BC)
       _bs_8 |  1000 -5.724212 -1.284417  4.947149  -15.43221   3.983783   (N)
             |                                      -17.34801     1.1554   (P)
             |                                      -14.55451   4.062553  (BC)
       _bs_9 |  1000 -.3162725 -.0418348  .2926239  -.8905005   .2579555   (N)
             |                                      -.9562694   .1152596   (P)
             |                                      -.8067846   .2161044  (BC)
      _bs_10 |  1000  2.197286  .4791749  1.435038  -.6187485   5.013321   (N)
             |                                       .6782664   5.474633   (P)
             |                                       .0262319   4.581345  (BC)
      _bs_11 |  1000  1.827374  .3449113  1.281948  -.6882453   4.342993   (N)
             |                                       .1559759   5.319251   (P)
             |                                      -.3702189   4.251153  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 1234

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype" "_b[
> wartype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _b[
> treaty] _b[untype] ", reps(1000) saving(C:\Temp\boot_untype) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unty
> pe
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[untype]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.037723 -.3101201  1.165435  -4.324704   .2492588   (N)
             |                                      -5.079943  -.6517755   (P)
             |                                       -4.28202  -.4280978  (BC)
       _bs_2 |  1000  -.550977 -.1185075  .3085761  -1.156509   .0545546   (N)
             |                                      -1.313139  -.2811089   (P)
             |                                      -1.022823  -.1763655  (BC)
       _bs_3 |  1000  .0089773   .002115  .0085404   -.007782   .0257365   (N)
             |                                      -.0005507   .0341465   (P)
             |                                      -.0017966   .0288976  (BC)
       _bs_4 |  1000 -1.759218 -.5517113   1.93005  -5.546635   2.028199   (N)
             |                                      -6.002833  -.0420855   (P)
             |                                      -4.055678   1.730385  (BC)
       _bs_5 |  1000  .1104262  .0415266  .2513656   -.382839   .6036913   (N)
             |                                      -.1456711   .5434556   (P)
             |                                      -.4015168   .3586739  (BC)
       _bs_6 |  1000  .0040713   .001855  .0060005  -.0077037   .0158463   (N)
             |                                      -.0005119   .0223684   (P)
             |                                      -.0024135   .0122653  (BC)
       _bs_7 |  1000  .0005114  .0002009  .0006157  -.0006968   .0017195   (N)
             |                                      -.0002172   .0021825   (P)
             |                                      -.0004464   .0015806  (BC)
       _bs_8 |  1000 -5.299446 -1.148194  4.988667  -15.08891   4.490022   (N)
             |                                      -16.86117   2.186427   (P)
             |                                      -14.38847   4.373326  (BC)
       _bs_9 |  1000 -.3015157 -.0426884   .284069  -.8589561   .2559246   (N)
             |                                      -.9203494   .1241695   (P)
             |                                      -.7596976   .2034339  (BC)
      _bs_10 |  1000   2.10146  .4216606  1.451944  -.7477501    4.95067   (N)
             |                                       .3954839   5.295391   (P)
             |                                       .1029555   4.380455  (BC)
      _bs_11 |  1000   .635233  .1416103   .420612  -.1901512   1.460617   (N)
             |                                       .1856772   1.671968   (P)
             |                                       .0464481   1.314984  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 1234

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty UNops" "_b[w
> artype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _b[t
> reaty] _b[UNops] ", reps(1000) saving(C:\Temp\boot_UNops) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty UNop
> s
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[UNops]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.059166 -.3117259  1.159236  -4.333983   .2156512   (N)
             |                                      -5.059483  -.6807868   (P)
             |                                      -4.215192  -.4499325  (BC)
       _bs_2 |  1000 -.5434793 -.1128207  .2917571  -1.116006   .0290477   (N)
             |                                      -1.259957  -.2753422   (P)
             |                                      -1.013694  -.1839027  (BC)
       _bs_3 |  1000  .0090216  .0020438  .0083234  -.0073117    .025355   (N)
             |                                      -.0003882   .0335583   (P)
             |                                      -.0015671   .0287715  (BC)
       _bs_4 |  1000 -1.651311 -.5137809  1.885094  -5.350508   2.047887   (N)
             |                                      -5.616938   .0604529   (P)
             |                                      -3.953382   1.846763  (BC)
       _bs_5 |  1000   .102684  .0390246  .2500368  -.3879735   .5933416   (N)
             |                                      -.1528355   .5146591   (P)
             |                                      -.4148193   .3509403  (BC)
       _bs_6 |  1000  .0037576  .0017598  .0059785  -.0079743   .0154896   (N)
             |                                      -.0008163   .0223641   (P)
             |                                      -.0023224   .0121063  (BC)
       _bs_7 |  1000  .0004996  .0001928  .0005925  -.0006631   .0016624   (N)
             |                                       -.000206   .0020402   (P)
             |                                      -.0004636   .0015804  (BC)
       _bs_8 |  1000  -5.63708 -1.231228  4.971716  -15.39328   4.119125   (N)
             |                                      -17.71975   1.759192   (P)
             |                                      -14.93738   3.943673  (BC)
       _bs_9 |  1000 -.2983887 -.0393962  .2740846  -.8362363   .2394589   (N)
             |                                      -.9252529    .116024   (P)
             |                                      -.7667538   .2051539  (BC)
      _bs_10 |  1000  2.195909  .4257307  1.382705  -.5174313   4.909248   (N)
             |                                       .5962212   5.380243   (P)
             |                                      -.0727525   4.434889  (BC)
      _bs_11 |  1000  .9922957  .2117928  .6790132    -.34016   2.324751   (N)
             |                                        .213616    2.65411   (P)
             |                                       .0186417   2.198352  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. * UNops combines mediation and observer missions (17 cases); traditional and multidimensional peacekeepi
> ng (15)
. 
. 
. ** use different seed
. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4" "_b
> [wartype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _b
> [treaty] _b[untype4] ", reps(1000) saving(C:\Temp\boot_multipko2) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unty
> pe4
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[untype4]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -1.742004 -.1950871    1.0466  -3.795791   .3117818   (N)
             |                                        -4.1013  -.3039747   (P)
             |                                       -3.75808  -.0289698  (BC)
       _bs_2 |  1000 -.4448499  -.090325  .2381606  -.9122023   .0225024   (N)
             |                                      -1.119416  -.1901672   (P)
             |                                      -.8119186  -.0709696  (BC)
       _bs_3 |  1000  .0063857  .0014202  .0073971  -.0081299   .0209014   (N)
             |                                      -.0034306    .024411   (P)
             |                                      -.0048093   .0212022  (BC)
       _bs_4 |  1000 -1.258933  .1583949  3.368568  -7.869214   5.351347   (N)
             |                                      -4.459064   2.692369   (P)
             |                                      -4.381945   2.777016  (BC)
       _bs_5 |  1000  .0616696 -.0701026  .6627037  -1.238781   1.362121   (N)
             |                                      -.5503153   .4001078   (P)
             |                                       -.488251   .4615749  (BC)
       _bs_6 |  1000  .0040934  .0026586  .0079316  -.0114711    .019658   (N)
             |                                       .0000562   .0276342   (P)
             |                                       -.002452     .01461  (BC)
       _bs_7 |  1000  .0007901  .0001876   .000922  -.0010192   .0025993   (N)
             |                                      -.0001783    .002485   (P)
             |                                      -.0003413   .0021717  (BC)
       _bs_8 |  1000 -6.016179 -1.542439  4.779912    -15.396   3.363641   (N)
             |                                      -18.22672   .1912315   (P)
             |                                      -14.23414   3.318699  (BC)
       _bs_9 |  1000 -.2987758 -.0400364  .2491315  -.7876569   .1901052   (N)
             |                                      -.8418344   .1323124   (P)
             |                                      -.7402839   .1643236  (BC)
      _bs_10 |  1000  2.124368  .4764076  1.310378    -.44704   4.695776   (N)
             |                                       .4966287   5.727833   (P)
             |                                      -.3396356   4.014587  (BC)
      _bs_11 |   633  3.135184  .1688494  1.745855  -.2931941   6.563563   (N)
             |                                       .7720723   7.042531   (P)
             |                                       .7401089   6.475263  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn" "_
> b[wartype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _
> b[treaty] _b[unintrvn] ", reps(1000) saving(C:\Temp\boot_unintrvn2) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unin
> trvn
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[unintrvn]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.036098 -.3370603  1.198043  -4.387068    .314872   (N)
             |                                      -5.158036  -.6888171   (P)
             |                                      -4.117395  -.2971287  (BC)
       _bs_2 |  1000 -.5545111 -.1166076  .3011433  -1.145457   .0364348   (N)
             |                                      -1.350419  -.2569347   (P)
             |                                      -1.069568  -.1708448  (BC)
       _bs_3 |  1000  .0087491  .0018013  .0074837  -.0059364   .0234346   (N)
             |                                      -.0007057   .0283093   (P)
             |                                      -.0017414     .02552  (BC)
       _bs_4 |  1000 -1.648769  -.638984  1.760838  -5.104134   1.806595   (N)
             |                                      -6.177068   .1216217   (P)
             |                                      -3.887495   1.649364  (BC)
       _bs_5 |  1000  .1056245  .0595326  .1919388  -.2710248   .4822739   (N)
             |                                      -.0999085   .5886967   (P)
             |                                      -.3516213   .3458254  (BC)
       _bs_6 |  1000  .0037728   .001673  .0057357  -.0074827   .0150282   (N)
             |                                      -.0008134   .0234108   (P)
             |                                      -.0021637   .0168131  (BC)
       _bs_7 |  1000  .0005694  .0001523  .0006235  -.0006542    .001793   (N)
             |                                      -.0002673   .0020801   (P)
             |                                      -.0004626   .0017522  (BC)
       _bs_8 |  1000 -5.724212 -1.481143  4.651951  -14.85293   3.404504   (N)
             |                                      -17.05118   .5370746   (P)
             |                                      -13.61728   3.858791  (BC)
       _bs_9 |  1000 -.3162725 -.0550226  .2589314  -.8243843   .1918393   (N)
             |                                      -.9468275    .122206   (P)
             |                                      -.7480544   .2263981  (BC)
      _bs_10 |  1000  2.197286  .5347718  1.337033  -.4264284   4.821001   (N)
             |                                       .6145144   6.211329   (P)
             |                                      -.3015265   4.092817  (BC)
      _bs_11 |  1000  1.827374  .3496141  1.218492   -.563724   4.218472   (N)
             |                                       .3165495   5.137056   (P)
             |                                      -.0575323   4.269723  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype" "_b[
> wartype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _b[
> treaty] _b[untype] ", reps(1000) saving(C:\Temp\boot_untype2) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unty
> pe
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[untype]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.037723 -.3474032  1.220733  -4.433217   .3577715   (N)
             |                                      -5.269595  -.7581314   (P)
             |                                      -3.991663  -.1496057  (BC)
       _bs_2 |  1000  -.550977 -.1185976  .3079147  -1.155211   .0532567   (N)
             |                                      -1.392085   -.268603   (P)
             |                                       -1.00424  -.1750236  (BC)
       _bs_3 |  1000  .0089773  .0019906  .0081369  -.0069901   .0249446   (N)
             |                                      -.0001232   .0296258   (P)
             |                                       -.002437   .0248467  (BC)
       _bs_4 |  1000 -1.759218 -.6165141  1.847296  -5.384243   1.865807   (N)
             |                                      -5.914268   .0663765   (P)
             |                                      -4.024189   1.791553  (BC)
       _bs_5 |  1000  .1104262  .0534121  .1947447  -.2717294   .4925817   (N)
             |                                      -.1217375   .5522144   (P)
             |                                      -.3301117   .3714072  (BC)
       _bs_6 |  1000  .0040713  .0018846  .0060041  -.0077109   .0158534   (N)
             |                                      -.0001869     .02499   (P)
             |                                      -.0016725   .0170984  (BC)
       _bs_7 |  1000  .0005114  .0001612  .0006534  -.0007709   .0017937   (N)
             |                                      -.0003127   .0019358   (P)
             |                                      -.0005557   .0015258  (BC)
       _bs_8 |  1000 -5.299446 -1.312602  4.854451  -14.82554   4.226645   (N)
             |                                      -16.54333   1.679887   (P)
             |                                      -13.75803   4.173745  (BC)
       _bs_9 |  1000 -.3015157 -.0580274  .2567682  -.8053827   .2023512   (N)
             |                                      -.9071399   .1256183   (P)
             |                                      -.7257605   .2153496  (BC)
      _bs_10 |  1000   2.10146  .4884174  1.363124  -.5734556   4.776376   (N)
             |                                        .315316   5.911843   (P)
             |                                      -.4284028   4.100825  (BC)
      _bs_11 |  1000   .635233  .1500554  .4178626  -.1847561   1.455222   (N)
             |                                       .1980897   1.739403   (P)
             |                                       .0902814    1.38502  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty UNops" "_b[w
> artype] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[decade] _b[t
> reaty] _b[UNops] ", reps(1000) saving(C:\Temp\boot_UNops2) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty UNop
> s
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[decade]
              _bs_10     = _b[treaty]
              _bs_11     = _b[UNops]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.059166 -.3471588  1.169761  -4.354636   .2363048   (N)
             |                                      -5.214713  -.7620504   (P)
             |                                      -3.981309  -.1556498  (BC)
       _bs_2 |  1000 -.5434793 -.1142123  .2867191   -1.10612   .0191614   (N)
             |                                      -1.355825  -.2685647   (P)
             |                                      -.9921554  -.1722915  (BC)
       _bs_3 |  1000  .0090216  .0019586  .0077601  -.0062064   .0242497   (N)
             |                                      -.0002585   .0292698   (P)
             |                                      -.0023301   .0244822  (BC)
       _bs_4 |  1000 -1.651311 -.5797405  1.680121  -4.948281    1.64566   (N)
             |                                      -6.000148   .1254658   (P)
             |                                       -3.86711   1.886206  (BC)
       _bs_5 |  1000   .102684  .0506669  .1796145  -.2497809    .455149   (N)
             |                                      -.1203372   .5491606   (P)
             |                                      -.3206097   .3615978  (BC)
       _bs_6 |  1000  .0037576  .0017971  .0059906   -.007998   .0155133   (N)
             |                                      -.0012391   .0250479   (P)
             |                                      -.0021729   .0192773  (BC)
       _bs_7 |  1000  .0004996   .000151  .0006192  -.0007154   .0017146   (N)
             |                                      -.0003332   .0019005   (P)
             |                                      -.0005676    .001467  (BC)
       _bs_8 |  1000  -5.63708 -1.376432   4.78124   -15.0195   3.745345   (N)
             |                                      -17.76865   1.178385   (P)
             |                                      -14.36862    4.49688  (BC)
       _bs_9 |  1000 -.2983887 -.0565787  .2508639  -.7906693   .1938919   (N)
             |                                      -.8888622   .1378812   (P)
             |                                      -.7232774   .1956214  (BC)
      _bs_10 |  1000  2.195909  .4897931  1.324789  -.4037789   4.795596   (N)
             |                                       .5131912   5.939337   (P)
             |                                      -.1577569   4.186186  (BC)
      _bs_11 |  1000  .9922957  .2385077  .6663006  -.3152136   2.299805   (N)
             |                                       .2488453   2.826354   (P)
             |                                      -.0270063   2.179955  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. 
. ** drop "decade" (LR test Prob > chi2 =     0.1202)
. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty untype4" "_b[wartyp
> e] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[treaty] _b[untype
> 4] ", reps(1000) saving(C:\Temp\boot_multipko3) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty untype4
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[treaty]
              _bs_10     = _b[untype4]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -1.876868 -.1801839  .9709601  -3.782223   .0284873   (N)
             |                                      -4.121824  -.5259423   (P)
             |                                      -3.829068  -.2582194  (BC)
       _bs_2 |  1000 -.4433597 -.0818294   .219628  -.8743448  -.0123747   (N)
             |                                      -1.037648  -.2093036   (P)
             |                                      -.8187971  -.0900086  (BC)
       _bs_3 |  1000  .0064033  .0012309  .0067444  -.0068315   .0196382   (N)
             |                                      -.0033237    .023172   (P)
             |                                      -.0046649    .019155  (BC)
       _bs_4 |  1000 -1.324481  .0994907  3.305132  -7.810278   5.161316   (N)
             |                                      -4.351989   2.134775   (P)
             |                                      -4.294753   2.289583  (BC)
       _bs_5 |  1000  .0703209 -.0574507  .6355311  -1.176808    1.31745   (N)
             |                                       -.461308   .4136426   (P)
             |                                       -.430252   .4403566  (BC)
       _bs_6 |  1000  .0045378    .00228  .0067105  -.0086305   .0177061   (N)
             |                                       .0001851   .0263741   (P)
             |                                      -.0020611    .014792  (BC)
       _bs_7 |  1000  .0007309  .0001771  .0007588  -.0007581   .0022198   (N)
             |                                      -.0002067   .0024149   (P)
             |                                      -.0003207   .0021167  (BC)
       _bs_8 |  1000 -6.340175  -1.22531   4.49994  -15.17059   2.490244   (N)
             |                                      -17.57272   .2057903   (P)
             |                                      -13.98824   3.387283  (BC)
       _bs_9 |  1000  1.808877  .4169203   1.18248   -.511552   4.129306   (N)
             |                                       .2300933   4.829698   (P)
             |                                      -.4320654   3.691906  (BC)
      _bs_10 |   633  3.027161  .0981238  1.700969  -.3130745   6.367396   (N)
             |                                       .5942515   6.758011   (P)
             |                                       .6210062   6.901234  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty unintrvn" "_b[warty
> pe] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[treaty] _b[unint
> rvn] ", reps(1000) saving(C:\Temp\boot_unintrvn3) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty unintrvn
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[treaty]
              _bs_10     = _b[unintrvn]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.189326 -.3167143  1.145232  -4.436661   .0580092   (N)
             |                                      -5.139894  -.9272257   (P)
             |                                      -4.216261   -.524275  (BC)
       _bs_2 |  1000 -.5434498 -.1060971  .2714761  -1.076179   -.010721   (N)
             |                                      -1.294436  -.2619884   (P)
             |                                      -.9983856  -.1606959  (BC)
       _bs_3 |  1000  .0087279  .0016682  .0068937  -.0047999   .0222556   (N)
             |                                      -.0003633   .0265616   (P)
             |                                      -.0016577   .0235601  (BC)
       _bs_4 |  1000 -1.698421 -.6125345  1.682912  -5.000868   1.604026   (N)
             |                                      -6.111344  -.0930479   (P)
             |                                      -3.897072   1.736544  (BC)
       _bs_5 |  1000  .1129213  .0579798  .1837352  -.2476299   .4734726   (N)
             |                                      -.0799924   .5945888   (P)
             |                                      -.3159102   .3547959  (BC)
       _bs_6 |  1000  .0042834  .0015626  .0055787   -.006664   .0152307   (N)
             |                                      -.0005521   .0224619   (P)
             |                                      -.0019577   .0157439  (BC)
       _bs_7 |  1000  .0005146  .0001533  .0005869  -.0006371   .0016664   (N)
             |                                      -.0002497   .0020154   (P)
             |                                      -.0003994   .0016569  (BC)
       _bs_8 |  1000 -6.178193 -1.163259   4.37664  -14.76665   2.410268   (N)
             |                                      -16.56816   .5165013   (P)
             |                                      -13.28796   2.176854  (BC)
       _bs_9 |  1000  1.838172  .4533548  1.176838  -.4711857    4.14753   (N)
             |                                       .4674726   4.792157   (P)
             |                                      -.4335741   3.405765  (BC)
      _bs_10 |  1000  1.755481  .2679963  1.167772  -.5360861   4.047048   (N)
             |                                       .2031946   4.873117   (P)
             |                                       .0107247   4.237525  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty untype" "_b[wartype
> ] _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[treaty] _b[untype]
>  ", reps(1000) saving(C:\Temp\boot_untype3) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty untype
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[treaty]
              _bs_10     = _b[untype]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.193852 -.3172316  1.123849  -4.399227   .0115225   (N)
             |                                       -5.18049  -.9528476   (P)
             |                                      -4.124817  -.4729267  (BC)
       _bs_2 |  1000 -.5417693 -.1074576  .2690312    -1.0697  -.0138383   (N)
             |                                      -1.310437  -.2632068   (P)
             |                                      -.9511395  -.1681408  (BC)
       _bs_3 |  1000  .0088752  .0017915   .007177  -.0052085    .022959   (N)
             |                                      -.0002405   .0267587   (P)
             |                                      -.0018926   .0229031  (BC)
       _bs_4 |  1000 -1.806697 -.5862036  1.688592  -5.120292   1.506898   (N)
             |                                      -5.947685  -.1171147   (P)
             |                                      -4.081875    1.26743  (BC)
       _bs_5 |  1000  .1174011  .0517293  .1798895  -.2356034   .4704056   (N)
             |                                      -.0951314   .5602274   (P)
             |                                      -.2872475   .3591008  (BC)
       _bs_6 |  1000  .0045051  .0017488  .0056693    -.00662   .0156301   (N)
             |                                       -.000233   .0240436   (P)
             |                                      -.0019214   .0150922  (BC)
       _bs_7 |  1000   .000487  .0001632  .0005992  -.0006889   .0016629   (N)
             |                                       -.000269   .0019412   (P)
             |                                      -.0005117   .0015073  (BC)
       _bs_8 |  1000 -5.780053 -1.027468  4.548055  -14.70489   3.144785   (N)
             |                                      -16.51575   1.512489   (P)
             |                                      -13.77293   3.665756  (BC)
       _bs_9 |  1000  1.750209  .4188325   1.18989   -.584761   4.085179   (N)
             |                                       .1800257   4.845586   (P)
             |                                      -.5709587   3.408474  (BC)
      _bs_10 |  1000  .6313336  .1221667  .3801739  -.1146974   1.377365   (N)
             |                                       .1698462   1.676323   (P)
             |                                       .0907822   1.385555  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. set seed 123456789

. 
. bs "logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty UNops" "_b[wartype]
>  _b[logcost] _b[wardur] _b[factnum] _b[factnum2] _b[trnsfcap] _b[develop] _b[exp] _b[treaty] _b[UNops] "
> , reps(1000) saving(C:\Temp\boot_UNops3) replace

command:      logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty UNops
statistics:   _bs_1      = _b[wartype]
              _bs_2      = _b[logcost]
              _bs_3      = _b[wardur]
              _bs_4      = _b[factnum]
              _bs_5      = _b[factnum2]
              _bs_6      = _b[trnsfcap]
              _bs_7      = _b[develop]
              _bs_8      = _b[exp]
              _bs_9      = _b[treaty]
              _bs_10     = _b[UNops]

Bootstrap statistics                              Number of obs    =       122
                                                  Replications     =      1000

------------------------------------------------------------------------------
Variable     |  Reps  Observed      Bias  Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  1000 -2.203248 -.3169136  1.095744  -4.353472  -.0530237   (N)
             |                                      -5.063879  -.9383176   (P)
             |                                      -4.132764  -.4217556  (BC)
       _bs_2 |  1000 -.5351756 -.1039223  .2592358  -1.043885  -.0264665   (N)
             |                                      -1.229069  -.2632948   (P)
             |                                      -.9361048  -.1762392  (BC)
       _bs_3 |  1000  .0089298  .0017665  .0070489  -.0049025   .0227622   (N)
             |                                      -.0002005   .0267834   (P)
             |                                      -.0019977    .022662  (BC)
       _bs_4 |  1000 -1.698897 -.5656826  1.620679  -4.879224   1.481429   (N)
             |                                      -5.861658  -.0380976   (P)
             |                                        -3.8285   1.057443  (BC)
       _bs_5 |  1000  .1096836   .050491  .1746551  -.2330492   .4524165   (N)
             |                                      -.0923065    .542535   (P)
             |                                      -.2621969   .3514971  (BC)
       _bs_6 |  1000  .0041903  .0017025  .0057282  -.0070503    .015431   (N)
             |                                      -.0008035    .023714   (P)
             |                                      -.0022551   .0151555  (BC)
       _bs_7 |  1000  .0004739  .0001567  .0005855  -.0006751   .0016228   (N)
             |                                       -.000289   .0019492   (P)
             |                                      -.0005349    .001475  (BC)
       _bs_8 |  1000 -6.052799  -1.07164  4.511788  -14.90647   2.800869   (N)
             |                                      -17.26148   1.046374   (P)
             |                                       -13.8618   4.079652  (BC)
       _bs_9 |  1000  1.843028  .4199652  1.160977  -.4352054   4.121261   (N)
             |                                       .3348148   4.872671   (P)
             |                                      -.4203266   3.474957  (BC)
      _bs_10 |  1000  .9907929  .1952111  .6247686  -.2352163   2.216802   (N)
             |                                       .2276847   2.728626   (P)
             |                                       .0338135   2.139334  (BC)
------------------------------------------------------------------------------
Note:  N   = normal
       P   = percentile
       BC  = bias-corrected

. 
. 
. 
. ** In the ISQ piece we refer to results from our model using interaction terms
. ** We focus on the interaction between MultiPKO and the Log of Deaths and Displacements
. ** In the supplement, we also include results from many more interactions.
. ** results obtained as follows:
. 
. 
. * Create interactions with MultiPKO: 
. 
. gen un4type=untype4*wartype

. gen un4fac=untype4*factnum

. gen un4fac2=untype4*factnum2

. gen un4tre=untype4*treaty

. gen un4tra=untype4*trnsfcap

. gen un4dev=untype4*develop

. gen un4exp=untype4*exp

. gen un4dur=untype4*wardur

. gen un4cost=untype4*logcost

. gen un4dec=untype4*decade

. 
. 
. * Now try to estimate the model adding interactions:
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4fac un4tre un4tra un4dev un4exp un4cost un4dur un4fac2 un4dec, cluster(cluster) nolog
note: untype4 dropped due to collinearity
note: un4fac dropped due to collinearity
note: un4tre dropped due to collinearity
note: un4exp dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     un4type |  -22.12195          .        .       .            .           .
      un4tra |  -.0140077          .        .       .            .           .
      un4dev |   .0168189          .        .       .            .           .
     un4cost |  -4.581178          .        .       .            .           .
      un4dur |   .2407945          .        .       .            .           .
     un4fac2 |    -.28541          .        .       .            .           .
      un4dec |    13.2649          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. 
. * You should get the following notes:
. * note: untype4 dropped due to collinearity
. * note: un4fac dropped due to collinearity
. * note: un4tre dropped due to collinearity
. * note: un4exp dropped due to collinearity
. 
. 
. * We re-estimated the model using combinations of fewer interaction terms given the multicolinearity pro
> blem
. 
. * Estimate the model in high vs. low ranges of the IVs and see if it fits differently
. * High/low ranges defined above/below means for each variable 
. 
. sum 

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
     cluster |         0
      decade |       122     3.47541    1.432868          1          6
      treaty |       122    .2868852    .4541727          0          1
      wardur |       122    79.57377    93.44401          1        600
     wartype |       122    .6393443    .4821709          0          1
-------------+--------------------------------------------------------
    unintrvn |       122     .295082    .4579603          0          1
      un2int |       122    .3688525    .4844835          0          1
     un2cint |       122     .852459    1.244307          0          4
      untype |       122    .8688525    1.465588          0          5
     factnum |       122    3.336066    1.529925          2         11
-------------+--------------------------------------------------------
      pbs2s3 |       122     .352459     .479706          0          1
    trnsfcap |       122    58.66762    171.7552       -112   1272.469
     logcost |       122    11.90362     2.42062   6.907755   15.67181
     untype4 |       122     .057377    .2335207          0          1
       UNops |       122    .4836066    .8355095          0          3
-------------+--------------------------------------------------------
    factnum2 |       122    13.45082    14.81238          4        121
         exp |       122    .1230395    .0832908       .013       .508
     develop |       122    542.7804    856.3503         10       5387
      unwdur |       122    23.04098    56.53587          0        288
    un_count |       122     .704918    .4579603          0          1
-------------+--------------------------------------------------------
   un4_count |       122     .942623    .2335207          0          1
un_count_dur |       122    56.53279    90.34463          0        600
     un4wdur |       122         8.5    41.06259          0        288
un4wdur__c~t |       122    71.07377    90.90592          0        600
     un4type |       122    .0245902    .1555111          0          1
-------------+--------------------------------------------------------
      un4fac |       122    .2704918    1.206298          0          8
     un4fac2 |       122    1.516393    7.948422          0         64
      un4tre |       122     .057377    .2335207          0          1
      un4tra |       122    4.389784    32.55576          0   344.5972
      un4dev |       122    17.93055     87.2428          0        584
-------------+--------------------------------------------------------
      un4exp |       122    .0059936    .0265693          0   .1577578
      un4dur |       122         8.5    41.06259          0        288
     un4cost |       122    .7758564    3.179314          0   15.20181
      un4dec |       122    .2295082    .9516136          0          5

. 
. gen highcost = logcost

. replace highcost=1 if logcost>11.90 & logcost !=.
(73 real changes made)

. replace highcost=0 if logcost<=11.90 & logcost !=.
(49 real changes made)

. 
. gen highdur = wardur

. replace highdur=1 if wardur>79.57 & wardur !=.
(44 real changes made)

. replace highdur=0 if wardur<=79.57 & wardur !=.
(78 real changes made)

. 
. gen highfact = factnum

. replace highfact=1 if factnum>3.34 & factnum !=.
(45 real changes made)

. replace highfact=0 if factnum>3.34 & factnum !=.
(45 real changes made)

. 
. gen hightran = trnsfcap

. replace hightran=1 if trnsfcap>58.66 & trnsfcap !=.
(20 real changes made)

. replace hightran=0 if trnsfcap<=58.66 & trnsfcap !=.
(102 real changes made)

. 
. gen highdev = develop

. replace highdev=1 if develop>542.78 & develop !=.
(34 real changes made)

. replace highdev=0 if develop<=542.78 & develop !=.
(88 real changes made)

. 
. gen highexp = exp

. replace highexp=1 if exp>.123 & exp !=.
(54 real changes made)

. replace highexp=0 if exp<=.123 & exp !=.
(68 real changes made)

. 
. gen highdec = decade

. replace highdec=1 if decade>3.47 & decade !=.
(63 real changes made)

. replace highdec=0 if decade<=3.47 & decade !=.
(59 real changes made)

. 
.  
. sum highcost highdur highfact hightran highdev highexp highdec

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
    highcost |       122    .5983607    .4922513          0          1
     highdur |       122    .3606557    .4821709          0          1
    highfact |       122    1.516393    1.228004          0          3
    hightran |       122    .1639344    .3717427          0          1
     highdev |       122    .2786885    .4502028          0          1
-------------+--------------------------------------------------------
     highexp |       122     .442623    .4987452          0          1
     highdec |       122    .5163934    .5017919          0          1

. 
. 
. 
. * Create interactions for LR tests
.  
. gen costtype = highcost * wartype

. gen costfac = highcost * factnum

. gen costfac2 = highcost * factnum2

. gen costtre = highcost * treaty

. gen costtra = highcost * trnsfcap

. gen costdev = highcost * develop

. gen costexp = highcost * exp

. gen costdur = highcost * wardur

. gen costdec = highcost * decade

. gen costun4 = highcost * untype4

. 
. gen durtype = highdur* wartype

. gen durfac = highdur* factnum

. gen durfac2 = highdur* factnum2

. gen durtre = highdur* treaty

. gen durtra = highdur* trnsfcap

. gen durdev = highdur* develop

. gen durexp = highdur* exp

. gen durcost = highdur* logcost

. gen durdec = highdur* decade

. gen durun4 = highdur* untype4

. 
. gen facttype = highfact * wartype

. gen facttre = highfact * treaty

. gen facttra = highfact * trnsfcap

. gen factdev = highfact * develop

. gen factexp = highfact * exp

. gen factdur = highfact * wardur

. gen factdec = highfact * decade

. gen factun4 = highfact * untype4

. gen factcost = highfact* logcost

. 
. gen trantype = hightran * wartype

. gen tranfac = hightran * factnum

. gen tranfac2 = hightran * factnum2

. gen trantre = hightran * treaty

. gen trandev = hightran * develop

. gen tranexp = hightran * exp

. gen trandur = hightran * wardur

. gen trandec = hightran * decade

. gen tranun4 = hightran * untype4

. gen trancost = hightran * logcost

. 
. gen devtype = highdev * wartype

. gen devfac = highdev * factnum

. gen devfac2 = highdev * factnum2

. gen devtre = highdev * treaty

. gen devtra = highdev * trnsfcap

. gen devexp = highdev * exp

. gen devdur = highdev * wardur

. gen devdec = highdev * decade

. gen devun4 = highdev * untype4

. gen devcost = highdev * logcost

. 
. gen exptype = highexp * wartype

. gen expfac = highexp * factnum

. gen expfac2 = highexp * factnum2

. gen exptre = highexp * treaty

. gen exptra = highexp * trnsfcap

. gen expdev = highexp * develop

. gen expdur = highexp * wardur

. gen expdec = highexp * decade

. gen expun4 = highexp * untype4

. gen expcost = highexp * logcost

. 
. gen dectype = highdec * wartype

. gen decfac = highdec * factnum

. gen decfac2 = highdec * factnum2

. gen dectre = highdec * treaty

. gen dectra = highdec * trnsfcap

. gen decdev = highdec * develop

. gen decexp = highdec * exp

. gen decdur = highdec * wardur

. gen decun4 = highdec * untype4

. gen deccost = highdec * logcost

. 
. gen tretype = treaty * wartype

. gen trefac = treaty * factnum

. gen trefac2 = treaty * factnum2

. gen tretra = treaty * trnsfcap

. gen tredev = treaty * develop

. gen treexp = treaty * exp

. gen tredur = treaty * wardur

. gen tredec = treaty * decade

. gen treun4 = treaty * untype4

. gen trecost = treaty * logcost

. 
. gen typetre = wartype * treaty

. gen typefac = wartype * factnum

. gen typefac2 = wartype * factnum2

. gen typetra = wartype * trnsfcap

. gen typedev = wartype * develop

. gen typeexp = wartype * exp

. gen typedur = wartype * wardur

. gen typedec = wartype * decade

. gen typeun4 = wartype * untype4

. gen typecost = wartype * logcost

. 
. * Interactions between war duration and all other covariates
.  
. gen dur_1  = wardur * treaty

. gen dur_2  = wardur * factnum

. gen dur_3  = wardur * factnum2

. gen dur_4  = wardur * trnsfcap

. gen dur_5  = wardur * develop

. gen dur_6  = wardur * exp

. gen dur_7  = wardur * wartype

. gen dur_8  = wardur * decade

. gen dur_9  = wardur * untype4

. gen dur_10  = wardur * logcost

. 
. * Interactions between logcost and all other covariates
.  
. gen cost_1  = logcost * treaty

. gen cost_2  = logcost * factnum

. gen cost_3  = logcost * factnum2

. gen cost_4  = logcost * trnsfcap

. gen cost_5  = logcost * develop

. gen cost_6  = logcost * exp

. gen cost_7  = logcost * wartype

. gen cost_8  = logcost * decade

. gen cost_9  = logcost * untype4

. gen cost_10  = logcost * wardur 

. 
. * Interactions between develop and all other covariates
.  
. gen dev_1  = develop * treaty

. gen dev_2  = develop * factnum

. gen dev_3  = develop * factnum2

. gen dev_4  = develop * trnsfcap

. gen dev_5  = develop * logcost

. gen dev_6  = develop * exp

. gen dev_7  = develop * wartype

. gen dev_8  = develop * decade

. gen dev_9  = develop * untype4

. gen dev_10  = develop * wardur 

.  
. * Interactions between exp and all other covariates
.  
. gen exp_1  = exp * treaty

. gen exp_2  = exp * factnum

. gen exp_3  = exp * factnum2

. gen exp_4  = exp * trnsfcap

. gen exp_5  = exp * logcost

. gen exp_6  = exp * develop

. gen exp_7  = exp * wartype

. gen exp_8  = exp * decade

. gen exp_9  = exp * untype4

. gen exp_10  = exp * wardur 

.  
. * Interactions between trnsfcap and all other covariates
.  
. gen tran_1  = trnsfcap * treaty

. gen tran_2  = trnsfcap * factnum

. gen tran_3  = trnsfcap * factnum2

. gen tran_4  = trnsfcap * exp

. gen tran_5  = trnsfcap * logcost

. gen tran_6  = trnsfcap * develop

. gen tran_7  = trnsfcap * wartype

. gen tran_8  = trnsfcap * decade

. gen tran_9  = trnsfcap * untype4

. gen tran_10  = trnsfcap * wardur 

.  
. * Interactions between decade and all other covariates
.  
. gen dec_1  = decade * treaty

. gen dec_2  = decade * factnum

. gen dec_3  = decade * factnum2

. gen dec_4  = decade * exp

. gen dec_5  = decade * logcost

. gen dec_6  = decade * develop

. gen dec_7  = decade * wartype

. gen dec_8  = decade * trnsfcap 

. gen dec_9  = decade * untype4

. gen dec_10  = decade * wardur 

.  
. * Interactions between factnum and all other covariates
.  
. gen fact_1  = factnum * treaty

. gen fact_2  = factnum * logcost

. * no interaction with factnum2
. gen fact_4  = factnum * trnsfcap

. gen fact_5  = factnum * develop

. gen fact_6  = factnum * exp

. gen fact_7  = factnum * wartype

. gen fact_8  = factnum * decade

. gen fact_9  = factnum * untype4

. gen fact_10  = factnum * wardur 

. 
. 
. * Model with wartype interactions
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4   typetr
> e typefac typefac2 typetra typedev typeexp typedur typedec typeun4 typecost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -37.901633                Pseudo R2       =     0.5213

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   -.983456   4.517907    -0.22   0.828    -9.838391    7.871479
     logcost |  -.9476953     .58768    -1.61   0.107    -2.099527    .2041364
      wardur |   .0081988   .0094971     0.86   0.388    -.0104152    .0268127
     factnum |  -.3465716    2.35729    -0.15   0.883    -4.966775    4.273632
    factnum2 |   .0248913   .2536343     0.10   0.922    -.4722228    .5220054
    trnsfcap |   .0326997   .0124336     2.63   0.009     .0083302    .0570692
     develop |   .0022448   .0013256     1.69   0.090    -.0003533    .0048429
         exp |   .3654783   2.779319     0.13   0.895    -5.081888    5.812844
      decade |  -.3977263   .4166476    -0.95   0.340    -1.214341    .4188879
      treaty |   2.833252   2.536424     1.12   0.264    -2.138047    7.804551
     untype4 |   21.27423   1.425631    14.92   0.000     18.48004    24.06841
     typetre |  -.0727489   2.861808    -0.03   0.980     -5.68179    5.536292
     typefac |  -1.656614    2.60747    -0.64   0.525    -6.767162    3.453934
    typefac2 |   .0863333   .2692301     0.32   0.748    -.4413481    .6140147
     typetra |   -.028097   .0124469    -2.26   0.024    -.0524924   -.0037015
     typedev |  -.0018048   .0015013    -1.20   0.229    -.0047472    .0011377
     typeexp |  -10.00255   5.214961    -1.92   0.055    -20.22368    .2185868
     typedur |  -.0022678   .0083829    -0.27   0.787    -.0186979    .0141624
     typedec |   .2984142   .4846771     0.62   0.538    -.6515354    1.248364
     typeun4 |  -18.66592          .        .       .            .           .
    typecost |   .4928716   .6546655     0.75   0.452    -.7902492    1.775992
       _cons |   9.232822    2.94769     3.13   0.002     3.455454    15.01019
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. * Did not run t-test because typeun4 is dropped 
. 
. * Model with treaty interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4   tretyp
> e trefac trefac2 tretra tredev treexp tredur tredec treun4 trecost, cluster(cluster) nolog
note: treun4 dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =      37.63
                                                  Prob > chi2     =     0.0098
Log pseudo-likelihood = -42.548147                Pseudo R2       =     0.4626

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   -2.07762   .7438522    -2.79   0.005    -3.535544   -.6196966
     logcost |  -.4703346   .1660515    -2.83   0.005    -.7957895   -.1448797
      wardur |   .0108141   .0069016     1.57   0.117    -.0027128     .024341
     factnum |  -2.222683   1.217901    -1.83   0.068    -4.609726    .1643598
    factnum2 |   .1596153   .1574355     1.01   0.311    -.1489526    .4681833
    trnsfcap |   .0054034   .0035865     1.51   0.132    -.0016261    .0124329
     develop |    .001303   .0005781     2.25   0.024     .0001699     .002436
         exp |  -5.631427   4.338773    -1.30   0.194    -14.13527    2.872411
      decade |  -.2675683   .2236998    -1.20   0.232    -.7060119    .1708754
      treaty |   8.337272   12.72593     0.66   0.512    -16.60509    33.27964
     untype4 |   4.308599   2.610307     1.65   0.099    -.8075088    9.424707
     tretype |    1.13245   1.730988     0.65   0.513    -2.260223    4.525123
      trefac |   2.246234   1.994249     1.13   0.260    -1.662423     6.15489
     trefac2 |  -.2440399   .2519061    -0.97   0.333    -.7377668     .249687
      tretra |  -.0013938   .0042477    -0.33   0.743    -.0097191    .0069316
      tredev |  -.0010018    .000915    -1.09   0.274    -.0027952    .0007916
      treexp |  -.8594292   6.163781    -0.14   0.889    -12.94022    11.22136
      tredur |  -.0136375   .0084735    -1.61   0.108    -.0302453    .0029702
      tredec |  -.4095137   .6162509    -0.66   0.506    -1.617343    .7983159
     trecost |  -.5804245   .7929265    -0.73   0.464    -2.134532    .9736829
       _cons |   10.19737   3.102777     3.29   0.001     4.116039     16.2787
------------------------------------------------------------------------------

. * treun4 dropped due to colinearity
. test tretype trefac trefac2 tretra tredev treexp tredur tredec trecost

 ( 1)  tretype = 0
 ( 2)  trefac = 0
 ( 3)  trefac2 = 0
 ( 4)  tretra = 0
 ( 5)  tredev = 0
 ( 6)  treexp = 0
 ( 7)  tredur = 0
 ( 8)  tredec = 0
 ( 9)  trecost = 0

           chi2(  9) =    9.57
         Prob > chi2 =    0.3860

. 
. * Model with factnum interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4   fact_1
>  fact_2 fact_4 fact_5 fact_6 fact_7 fact_8 fact_9 fact_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =      40.31
                                                  Prob > chi2     =     0.0046
Log pseudo-likelihood = -37.532082                Pseudo R2       =     0.5259

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   2.796387   2.291801     1.22   0.222     -1.69546    7.288234
     logcost |  -.6780868   .3734738    -1.82   0.069    -1.410082    .0539084
      wardur |   -.019144    .012467    -1.54   0.125    -.0435789    .0052909
     factnum |   1.850449   3.237979     0.57   0.568    -4.495873    8.196772
    factnum2 |  -.3506068   .4693957    -0.75   0.455    -1.270606    .5693919
    trnsfcap |   .0267336   .0112181     2.38   0.017     .0047465    .0487208
     develop |   .0027152   .0016107     1.69   0.092    -.0004418    .0058722
         exp |   6.829217   11.57865     0.59   0.555    -15.86452    29.52296
      decade |   1.184459   .7702215     1.54   0.124    -.3251478    2.694065
      treaty |  -3.960288   2.949483    -1.34   0.179    -9.741169    1.820593
     untype4 |   4.893898   3.605261     1.36   0.175    -2.172285    11.96008
      fact_1 |     2.7369   1.320583     2.07   0.038     .1486039    5.325196
      fact_2 |   .0924859   .1376141     0.67   0.502    -.1772328    .3622045
      fact_4 |  -.0038534   .0021051    -1.83   0.067    -.0079794    .0002726
      fact_5 |  -.0006408   .0005951    -1.08   0.282    -.0018071    .0005256
      fact_6 |  -5.549197   5.093819    -1.09   0.276     -15.5329    4.434504
      fact_7 |  -1.515785    .929996    -1.63   0.103    -3.338544    .3069737
      fact_8 |  -.6148149   .3156797    -1.95   0.051    -1.233536    .0039059
      fact_9 |  -.4952846   .6914244    -0.72   0.474    -1.850451    .8598822
     fact_10 |   .0102887    .004512     2.28   0.023     .0014453     .019132
       _cons |   2.658873    5.60547     0.47   0.635    -8.327646    13.64539
------------------------------------------------------------------------------

note: 0 failures and 1 success completely determined.

. test fact_1 fact_2 fact_4 fact_5 fact_6 fact_7 fact_8 fact_9 fact_10

 ( 1)  fact_1 = 0
 ( 2)  fact_2 = 0
 ( 3)  fact_4 = 0
 ( 4)  fact_5 = 0
 ( 5)  fact_6 = 0
 ( 6)  fact_7 = 0
 ( 7)  fact_8 = 0
 ( 8)  fact_9 = 0
 ( 9)  fact_10 = 0

           chi2(  9) =   14.14
         Prob > chi2 =    0.1174

. 
. * Model with decade interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4   dec_1 
> dec_2 dec_3 dec_4 dec_5 dec_6 dec_7 dec_8 dec_9 dec_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -42.250349                Pseudo R2       =     0.4664

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.074126   1.162777    -1.78   0.074    -4.353126    .2048748
     logcost |  -.5252921   .3974095    -1.32   0.186      -1.3042    .2536163
      wardur |   .0219161   .0182068     1.20   0.229    -.0137685    .0576007
     factnum |  -2.756084   4.232591    -0.65   0.515    -11.05181    5.539643
    factnum2 |   .3400265   .6359502     0.53   0.593    -.9064131    1.586466
    trnsfcap |   .0121588   .0095055     1.28   0.201    -.0064716    .0307891
     develop |   .0010949   .0011484     0.95   0.340     -.001156    .0033458
         exp |  -1.136759   8.641076    -0.13   0.895    -18.07296    15.79944
      decade |  -.6458902   2.128212    -0.30   0.762    -4.817109    3.525328
      treaty |   2.712066   2.528694     1.07   0.283    -2.244085    7.668216
     untype4 |   104.2221          .        .       .            .           .
       dec_1 |  -.0439834   .5940869    -0.07   0.941    -1.208372    1.120406
       dec_2 |    .518169   1.336904     0.39   0.698    -2.102115    3.138453
       dec_3 |  -.0937018   .2077233    -0.45   0.652    -.5008319    .3134283
       dec_4 |   -1.77884   1.900166    -0.94   0.349    -5.503097    1.945418
       dec_5 |   .0198552   .1131208     0.18   0.861    -.2018574    .2415678
       dec_6 |  -.0001654   .0003137    -0.53   0.598    -.0007803    .0004495
       dec_7 |   .2386752   .3566527     0.67   0.503    -.4603511    .9377016
       dec_8 |   -.002725   .0029437    -0.93   0.355    -.0084944    .0030445
       dec_9 |  -20.70395   .3367715   -61.48   0.000    -21.36401   -20.04389
      dec_10 |  -.0056221   .0050139    -1.12   0.262    -.0154492     .004205
       _cons |   9.655726   7.298854     1.32   0.186    -4.649765    23.96122
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. test dec_1 dec_2 dec_3 dec_4 dec_5 dec_6 dec_7 dec_8 dec_9 dec_10

 ( 1)  dec_1 = 0
 ( 2)  dec_2 = 0
 ( 3)  dec_3 = 0
 ( 4)  dec_4 = 0
 ( 5)  dec_5 = 0
 ( 6)  dec_6 = 0
 ( 7)  dec_7 = 0
 ( 8)  dec_8 = 0
 ( 9)  dec_9 = 0
 (10)  dec_10 = 0

           chi2( 10) = 9583.25
         Prob > chi2 =    0.0000

. 
. * Model with exp interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4    exp_1
>  exp_2 exp_3 exp_4 exp_5 exp_6 exp_7 exp_8 exp_9 exp_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      58.98
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -42.650076                Pseudo R2       =     0.4613

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.104618   1.151799    -0.96   0.338    -3.362102    1.152866
     logcost |   .0286711   .2507324     0.11   0.909    -.4627552    .5200975
      wardur |   .0045229    .010842     0.42   0.677    -.0167271    .0257729
     factnum |   1.463601   2.397593     0.61   0.542    -3.235596    6.162798
    factnum2 |  -.2743748    .328416    -0.84   0.403    -.9180584    .3693088
    trnsfcap |   .0059219   .0063405     0.93   0.350    -.0065052     .018349
     develop |   .0007978   .0009698     0.82   0.411     -.001103    .0026986
         exp |   104.7869   66.10945     1.59   0.113    -24.78527     234.359
      decade |    .171701   .3558822     0.48   0.629    -.5258153    .8692173
      treaty |   1.437485   1.531596     0.94   0.348    -1.564389    4.439359
     untype4 |   3.428882   3.558187     0.96   0.335    -3.545037     10.4028
       exp_1 |    10.2914   13.74773     0.75   0.454    -16.65365    37.23645
       exp_2 |  -29.18276   27.30401    -1.07   0.285    -82.69764    24.33212
       exp_3 |   3.613807   3.717992     0.97   0.331    -3.673322    10.90094
       exp_4 |  -.0297788   .0770916    -0.39   0.699    -.1808756     .121318
       exp_5 |  -4.640053   2.427507    -1.91   0.056     -9.39788    .1177743
       exp_6 |   .0005966   .0107313     0.06   0.956    -.0204364    .0216295
       exp_7 |  -8.297942   9.149677    -0.91   0.364    -26.23098    9.635096
       exp_8 |  -3.698374   3.031319    -1.22   0.222    -9.639649    2.242902
       exp_9 |  -1.831788   29.24755    -0.06   0.950    -59.15593    55.49236
      exp_10 |   .0148509   .0728769     0.20   0.839    -.1279852     .157687
       _cons |  -2.663738   6.373795    -0.42   0.676    -15.15615    9.828671
------------------------------------------------------------------------------

. test exp_1 exp_2 exp_3 exp_4 exp_5 exp_6 exp_7 exp_8 exp_9 exp_10

 ( 1)  exp_1 = 0
 ( 2)  exp_2 = 0
 ( 3)  exp_3 = 0
 ( 4)  exp_4 = 0
 ( 5)  exp_5 = 0
 ( 6)  exp_6 = 0
 ( 7)  exp_7 = 0
 ( 8)  exp_8 = 0
 ( 9)  exp_9 = 0
 (10)  exp_10 = 0

           chi2( 10) =    6.53
         Prob > chi2 =    0.7690

. 
. * Model with develop interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4   dev_1 
> dev_2 dev_3 dev_4 dev_5 dev_6 dev_7 dev_8 dev_9 dev_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      63.05
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -39.531057                Pseudo R2       =     0.5007

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -.4733786   .9584799    -0.49   0.621    -2.351965    1.405208
     logcost |  -.1083993   .1715179    -0.63   0.527    -.4445682    .2277697
      wardur |   .0053779   .0071181     0.76   0.450    -.0085733    .0193291
     factnum |  -.6684621   .7062211    -0.95   0.344     -2.05263    .7157058
    factnum2 |   .0270566   .0600803     0.45   0.652    -.0906985    .1448117
    trnsfcap |   .0002413   .0049785     0.05   0.961    -.0095164    .0099991
     develop |   .0319574    .015816     2.02   0.043     .0009586    .0629561
         exp |  -11.76473   5.924854    -1.99   0.047    -23.37723   -.1522325
      decade |  -.1312461   .2414456    -0.54   0.587    -.6044707    .3419786
      treaty |   1.683302    .929168     1.81   0.070    -.1378334    3.504438
     untype4 |   5.685213   2.449201     2.32   0.020     .8848671    10.48556
       dev_1 |   .0037595    .002363     1.59   0.112    -.0008719    .0083909
       dev_2 |  -.0022942   .0024404    -0.94   0.347    -.0070773    .0024888
       dev_3 |   .0001643   .0002631     0.62   0.532    -.0003514    .0006801
       dev_4 |   .0000106   5.90e-06     1.80   0.072    -9.32e-07    .0000222
       dev_5 |  -.0018947   .0009005    -2.10   0.035    -.0036597   -.0001297
       dev_6 |   .0008644   .0112529     0.08   0.939    -.0211909    .0229197
       dev_7 |  -.0080533   .0034035    -2.37   0.018     -.014724   -.0013825
       dev_8 |   .0003471   .0004648     0.75   0.455    -.0005638     .001258
       dev_9 |  -.0043001   .0059381    -0.72   0.469    -.0159386    .0073383
      dev_10 |  -8.18e-06   9.14e-06    -0.90   0.371    -.0000261    9.73e-06
       _cons |    2.77232   2.569528     1.08   0.281    -2.263862    7.808503
------------------------------------------------------------------------------

note: 1 failure and 2 successes completely determined.

. test dev_1 dev_2 dev_3 dev_4 dev_5 dev_6 dev_7 dev_8 dev_9 dev_10

 ( 1)  dev_1 = 0
 ( 2)  dev_2 = 0
 ( 3)  dev_3 = 0
 ( 4)  dev_4 = 0
 ( 5)  dev_5 = 0
 ( 6)  dev_6 = 0
 ( 7)  dev_7 = 0
 ( 8)  dev_8 = 0
 ( 9)  dev_9 = 0
 (10)  dev_10 = 0

           chi2( 10) =   16.59
         Prob > chi2 =    0.0841

. 
. * Model with logcost interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4  cost_1 
> cost_2 cost_3 cost_4 cost_5 cost_6 cost_7 cost_8 cost_9 cost_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      46.23
                                                  Prob > chi2     =     0.0012
Log pseudo-likelihood = -33.151423                Pseudo R2       =     0.5813

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   -3.62562   3.189342    -1.14   0.256    -9.876616    2.625375
     logcost |  -6.223617   2.253048    -2.76   0.006    -10.63951   -1.807724
      wardur |   .0641325   .0570587     1.12   0.261    -.0477005    .1759655
     factnum |  -57.56304   21.98866    -2.62   0.009      -100.66   -14.46607
    factnum2 |   8.540664   3.475616     2.46   0.014     1.728582    15.35275
    trnsfcap |  -.0195663    .014894    -1.31   0.189    -.0487581    .0096254
     develop |   .0131487   .0060721     2.17   0.030     .0012476    .0250499
         exp |   57.50057   31.53741     1.82   0.068    -4.311616    119.3128
      decade |  -.0435522   .9314919    -0.05   0.963    -1.869243    1.782138
      treaty |    2.67373   4.953387     0.54   0.589     -7.03473    12.38219
     untype4 |   39.42959   37.28768     1.06   0.290    -33.65291    112.5121
      cost_1 |  -.0147704   .4058627    -0.04   0.971    -.8102467     .780706
      cost_2 |   4.553758   1.686875     2.70   0.007     1.247543    7.859973
      cost_3 |  -.6843293    .268488    -2.55   0.011    -1.210556   -.1581025
      cost_4 |   .0025574   .0013807     1.85   0.064    -.0001486    .0052635
      cost_5 |  -.0009779   .0004653    -2.10   0.036    -.0018899    -.000066
      cost_6 |  -5.816923    3.04093    -1.91   0.056    -11.77704    .1431907
      cost_7 |   .0981446    .323291     0.30   0.761    -.5354941    .7317833
      cost_8 |  -.0034431   .0928507    -0.04   0.970    -.1854271    .1785408
      cost_9 |   -2.22789    2.51297    -0.89   0.375    -7.153221    2.697441
     cost_10 |  -.0043652   .0042308    -1.03   0.302    -.0126575    .0039271
       _cons |    79.7579   29.29685     2.72   0.006     22.33714    137.1787
------------------------------------------------------------------------------

note: 3 failures and 2 successes completely determined.

. test cost_1 cost_2 cost_3 cost_4 cost_5 cost_6 cost_7 cost_8 cost_9 cost_10

 ( 1)  cost_1 = 0
 ( 2)  cost_2 = 0
 ( 3)  cost_3 = 0
 ( 4)  cost_4 = 0
 ( 5)  cost_5 = 0
 ( 6)  cost_6 = 0
 ( 7)  cost_7 = 0
 ( 8)  cost_8 = 0
 ( 9)  cost_9 = 0
 (10)  cost_10 = 0

           chi2( 10) =   16.06
         Prob > chi2 =    0.0979

. 
. * Model with trnsfcap interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4  tran_1 
> tran_2 tran_3 tran_4 tran_5 tran_6 tran_7 tran_8 tran_9 tran_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =     149.73
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -36.392147                Pseudo R2       =     0.5403

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   .1130718   .7606184     0.15   0.882    -1.377713    1.603857
     logcost |   -.675411   .1537467    -4.39   0.000    -.9767489   -.3740731
      wardur |   .0087494   .0075705     1.16   0.248    -.0060885    .0235873
     factnum |   .7630613    .824572     0.93   0.355    -.8530701    2.379193
    factnum2 |  -.1675687    .083901    -2.00   0.046    -.3320116   -.0031258
    trnsfcap |   .2175352   .0912986     2.38   0.017     .0385932    .3964771
     develop |   .0007562   .0005568     1.36   0.174    -.0003351    .0018475
         exp |  -6.299826   4.776653    -1.32   0.187    -15.66189    3.062241
      decade |  -.0587344   .2604581    -0.23   0.822    -.5692229    .4517541
      treaty |   2.293909   1.064488     2.15   0.031     .2075498    4.380267
     untype4 |   7.058296   2.324815     3.04   0.002     2.501743    11.61485
      tran_1 |   .0327441   .0445982     0.73   0.463    -.0546667     .120155
      tran_2 |  -.1123554   .0411178    -2.73   0.006    -.1929448    -.031766
      tran_3 |   .0111441   .0042653     2.61   0.009     .0027842     .019504
      tran_4 |   .0654655    .268821     0.24   0.808    -.4614139     .592345
      tran_5 |   .0099453   .0052665     1.89   0.059     -.000377    .0202675
      tran_6 |   .0000155   .0000201     0.77   0.440    -.0000239    .0000549
      tran_7 |  -.0658871   .0305219    -2.16   0.031     -.125709   -.0060652
      tran_8 |  -.0164835   .0088582    -1.86   0.063    -.0338453    .0008782
      tran_9 |  -.0487954   .0262898    -1.86   0.063    -.1003225    .0027316
     tran_10 |  -.0001212   .0001632    -0.74   0.458    -.0004411    .0001987
       _cons |    5.09148   2.391629     2.13   0.033      .403974    9.778987
------------------------------------------------------------------------------

note: 1 failure and 3 successes completely determined.

. test tran_1 tran_2 tran_3 tran_4 tran_5 tran_6 tran_7 tran_8 tran_9 tran_10

 ( 1)  tran_1 = 0
 ( 2)  tran_2 = 0
 ( 3)  tran_3 = 0
 ( 4)  tran_4 = 0
 ( 5)  tran_5 = 0
 ( 6)  tran_6 = 0
 ( 7)  tran_7 = 0
 ( 8)  tran_8 = 0
 ( 9)  tran_9 = 0
 (10)  tran_10 = 0

           chi2( 10) =   32.79
         Prob > chi2 =    0.0003

. 
. * Model with war duration interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4   dur_1 
>  dur_2 dur_3 dur_4 dur_5 dur_6 dur_7 dur_8 dur_9 dur_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      62.16
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -38.788107                Pseudo R2       =     0.5101

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.293308   .8285619    -2.77   0.006    -3.917259   -.6693564
     logcost |  -.2644656   .1965006    -1.35   0.178    -.6495998    .1206685
      wardur |   .0487587   .0454987     1.07   0.284    -.0404172    .1379346
     factnum |  -2.740188   2.315966    -1.18   0.237    -7.279398    1.799022
    factnum2 |   .1464593   .3161532     0.46   0.643    -.4731896    .7661082
    trnsfcap |   .0065515   .0025824     2.54   0.011     .0014901    .0116129
     develop |   .0007407   .0006214     1.19   0.233    -.0004772    .0019586
         exp |  -4.165366   5.407638    -0.77   0.441    -14.76414    6.433409
      decade |  -.1713125   .2390573    -0.72   0.474    -.6398563    .2972313
      treaty |   3.748681   1.316022     2.85   0.004     1.169326    6.328036
     untype4 |  -.4988225   1.700596    -0.29   0.769    -3.831929    2.834284
       dur_1 |  -.0117231   .0086851    -1.35   0.177    -.0287456    .0052995
       dur_2 |   .0207642   .0190157     1.09   0.275    -.0165059    .0580344
       dur_3 |  -.0016519    .002278    -0.73   0.468    -.0061167    .0028129
       dur_4 |  -.0000167   .0000172    -0.97   0.332    -.0000504     .000017
       dur_5 |  -9.24e-07   3.63e-06    -0.25   0.799    -8.03e-06    6.18e-06
       dur_6 |  -.0702249   .0776149    -0.90   0.366    -.2223474    .0818975
       dur_7 |   .0112858   .0080665     1.40   0.162    -.0045243    .0270959
       dur_8 |  -.0075411   .0044912    -1.68   0.093    -.0163437    .0012616
       dur_9 |   .0604008   .0148733     4.06   0.000     .0312497    .0895519
      dur_10 |  -.0048335    .002645    -1.83   0.068    -.0100175    .0003505
       _cons |   9.186757   4.713122     1.95   0.051    -.0507914    18.42431
------------------------------------------------------------------------------

. test dur_1 dur_2 dur_3 dur_4 dur_5 dur_6 dur_7 dur_8 dur_9 dur_10

 ( 1)  dur_1 = 0
 ( 2)  dur_2 = 0
 ( 3)  dur_3 = 0
 ( 4)  dur_4 = 0
 ( 5)  dur_5 = 0
 ( 6)  dur_6 = 0
 ( 7)  dur_7 = 0
 ( 8)  dur_8 = 0
 ( 9)  dur_9 = 0
 (10)  dur_10 = 0

           chi2( 10) =   23.65
         Prob > chi2 =    0.0086

. 
. * Same model, dropping decade because wars that started later will have lower war duration
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty untype4   dur_1 dur_2 d
> ur_3 dur_4 dur_5 dur_6 dur_7 dur_9 dur_10, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(19)   =      62.92
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood =  -41.82186                Pseudo R2       =     0.4718

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.277673   .8307098    -2.74   0.006    -3.905834   -.6495118
     logcost |  -.3159159   .1847712    -1.71   0.087    -.6780608     .046229
      wardur |  -.0120983   .0393207    -0.31   0.758    -.0891654    .0649688
     factnum |   -2.98571   2.282737    -1.31   0.191    -7.459792    1.488372
    factnum2 |   .1992292   .3147065     0.63   0.527    -.4175843    .8160426
    trnsfcap |   .0064997   .0025769     2.52   0.012      .001449    .0115503
     develop |   .0007454   .0006699     1.11   0.266    -.0005676    .0020585
         exp |  -6.598953   4.986461    -1.32   0.186    -16.37224     3.17433
      treaty |   2.993168   1.292257     2.32   0.021     .4603904    5.525945
     untype4 |  -.4575095   1.791198    -0.26   0.798    -3.968193    3.053174
       dur_1 |  -.0097233   .0090427    -1.08   0.282    -.0274467    .0080002
       dur_2 |   .0285698   .0187556     1.52   0.128    -.0081905    .0653301
       dur_3 |  -.0029497    .002249    -1.31   0.190    -.0073577    .0014582
       dur_4 |  -.0000148   .0000212    -0.70   0.484    -.0000564    .0000268
       dur_5 |   1.45e-06   4.44e-06     0.33   0.744    -7.26e-06    .0000102
       dur_6 |  -.0194678   .0657428    -0.30   0.767    -.1483212    .1093857
       dur_7 |   .0058818   .0073739     0.80   0.425    -.0085708    .0203345
       dur_9 |   .0525786   .0173018     3.04   0.002     .0186677    .0864896
      dur_10 |  -.0027657   .0019844    -1.39   0.163    -.0066551    .0011237
       _cons |   9.736988   4.687478     2.08   0.038     .5496997    18.92428
------------------------------------------------------------------------------

note: 1 failure and 0 successes completely determined.

. test dur_1 dur_2 dur_3 dur_4 dur_5 dur_6 dur_7 dur_9 dur_10

 ( 1)  dur_1 = 0
 ( 2)  dur_2 = 0
 ( 3)  dur_3 = 0
 ( 4)  dur_4 = 0
 ( 5)  dur_5 = 0
 ( 6)  dur_6 = 0
 ( 7)  dur_7 = 0
 ( 8)  dur_9 = 0
 ( 9)  dur_10 = 0

           chi2(  9) =   15.66
         Prob > chi2 =    0.0744

. 
. 
. ** Interactions with any UN intervention (unintrvn)
. ** We repeat the same excercise, now adding interactions with Unintrvn (instead of MultiPKO). 
. 
. use "C:\temp\DS2000replication.dta", clear

. 
. gen highcost = logcost

. replace highcost=1 if logcost>11.90 & logcost !=.
(73 real changes made)

. replace highcost=0 if logcost<=11.90 & logcost !=.
(49 real changes made)

. 
. gen highdur = wardur

. replace highdur=1 if wardur>79.57 & wardur !=.
(44 real changes made)

. replace highdur=0 if wardur<=79.57 & wardur !=.
(78 real changes made)

. 
. gen highfact = factnum

. replace highfact=1 if factnum>3.34 & factnum !=.
(45 real changes made)

. replace highfact=0 if factnum>3.34 & factnum !=.
(45 real changes made)

. 
. gen hightran = trnsfcap

. replace hightran=1 if trnsfcap>58.66 & trnsfcap !=.
(20 real changes made)

. replace hightran=0 if trnsfcap<=58.66 & trnsfcap !=.
(102 real changes made)

. 
. gen highdev = develop

. replace highdev=1 if develop>542.78 & develop !=.
(34 real changes made)

. replace highdev=0 if develop<=542.78 & develop !=.
(88 real changes made)

. 
. gen highexp = exp

. replace highexp=1 if exp>.123 & exp !=.
(54 real changes made)

. replace highexp=0 if exp<=.123 & exp !=.
(68 real changes made)

. 
. gen highdec = decade

. replace highdec=1 if decade>3.47 & decade !=.
(63 real changes made)

. replace highdec=0 if decade<=3.47 & decade !=.
(59 real changes made)

. 
. gen costtype = highcost * wartype

. gen costfac = highcost * factnum

. gen costfac2 = highcost * factnum2

. gen costtre = highcost * treaty

. gen costtra = highcost * trnsfcap

. gen costdev = highcost * develop

. gen costexp = highcost * exp

. gen costdur = highcost * wardur

. gen costdec = highcost * decade

. gen costun1 = highcost * unintrvn

. 
. gen durtype = highdur* wartype

. gen durfac = highdur* factnum

. gen durfac2 = highdur* factnum2

. gen durtre = highdur* treaty

. gen durtra = highdur* trnsfcap

. gen durdev = highdur* develop

. gen durexp = highdur* exp

. gen durcost = highdur* logcost

. gen durdec = highdur* decade

. gen durun1 = highdur* unintrvn

. 
. gen facttype = highfact * wartype

. gen facttre = highfact * treaty

. gen facttra = highfact * trnsfcap

. gen factdev = highfact * develop

. gen factexp = highfact * exp

. gen factdur = highfact * wardur

. gen factdec = highfact * decade

. gen factun1 = highfact * unintrvn

. gen factcost = highfact* logcost

. 
. gen trantype = hightran * wartype

. gen tranfac = hightran * factnum

. gen tranfac2 = hightran * factnum2

. gen trantre = hightran * treaty

. gen trandev = hightran * develop

. gen tranexp = hightran * exp

. gen trandur = hightran * wardur

. gen trandec = hightran * decade

. gen tranun1 = hightran * unintrvn

. gen trancost = hightran * logcost

. 
. gen devtype = highdev * wartype

. gen devfac = highdev * factnum

. gen devfac2 = highdev * factnum2

. gen devtre = highdev * treaty

. gen devtra = highdev * trnsfcap

. gen devexp = highdev * exp

. gen devdur = highdev * wardur

. gen devdec = highdev * decade

. gen devun1 = highdev * unintrvn

. gen devcost = highdev * logcost

. 
. gen exptype = highexp * wartype

. gen expfac = highexp * factnum

. gen expfac2 = highexp * factnum2

. gen exptre = highexp * treaty

. gen exptra = highexp * trnsfcap

. gen expdev = highexp * develop

. gen expdur = highexp * wardur

. gen expdec = highexp * decade

. gen expun1 = highexp * unintrvn

. gen expcost = highexp * logcost

. 
. gen dectype = highdec * wartype

. gen decfac = highdec * factnum

. gen decfac2 = highdec * factnum2

. gen dectre = highdec * treaty

. gen dectra = highdec * trnsfcap

. gen decdev = highdec * develop

. gen decexp = highdec * exp

. gen decdur = highdec * wardur

. gen decun1 = highdec * unintrvn

. gen deccost = highdec * logcost

. 
. gen tretype = treaty * wartype

. gen trefac = treaty * factnum

. gen trefac2 = treaty * factnum2

. gen tretra = treaty * trnsfcap

. gen tredev = treaty * develop

. gen treexp = treaty * exp

. gen tredur = treaty * wardur

. gen tredec = treaty * decade

. gen treun1 = treaty * unintrvn

. gen trecost = treaty * logcost

. 
. gen typetre = wartype * treaty

. gen typefac = wartype * factnum

. gen typefac2 = wartype * factnum2

. gen typetra = wartype * trnsfcap

. gen typedev = wartype * develop

. gen typeexp = wartype * exp

. gen typedur = wartype * wardur

. gen typedec = wartype * decade

. gen typeun1 = wartype * unintrvn

. gen typecost = wartype * logcost

. 
. * interactions between war duration and all other covariates
.  
. gen dur_1  = wardur * treaty

. gen dur_2  = wardur * factnum

. gen dur_3  = wardur * factnum2

. gen dur_4  = wardur * trnsfcap

. gen dur_5  = wardur * develop

. gen dur_6  = wardur * exp

. gen dur_7  = wardur * wartype

. gen dur_8  = wardur * decade

. gen dur_9  = wardur * unintrvn

. gen dur_10  = wardur * logcost

. 
. * interactions between logcost and all other covariates
.  
. gen cost_1  = logcost * treaty

. gen cost_2  = logcost * factnum

. gen cost_3  = logcost * factnum2

. gen cost_4  = logcost * trnsfcap

. gen cost_5  = logcost * develop

. gen cost_6  = logcost * exp

. gen cost_7  = logcost * wartype

. gen cost_8  = logcost * decade

. gen cost_9  = logcost * unintrvn

. gen cost_10  = logcost * wardur 

. 
. * interactions between develop and all other covariates
.  
. gen dev_1  = develop * treaty

. gen dev_2  = develop * factnum

. gen dev_3  = develop * factnum2

. gen dev_4  = develop * trnsfcap

. gen dev_5  = develop * logcost

. gen dev_6  = develop * exp

. gen dev_7  = develop * wartype

. gen dev_8  = develop * decade

. gen dev_9  = develop * unintrvn

. gen dev_10  = develop * wardur 

. 
. * interactions between exp and all other covariates
.  
. gen exp_1  = exp * treaty

. gen exp_2  = exp * factnum

. gen exp_3  = exp * factnum2

. gen exp_4  = exp * trnsfcap

. gen exp_5  = exp * logcost

. gen exp_6  = exp * develop

. gen exp_7  = exp * wartype

. gen exp_8  = exp * decade

. gen exp_9  = exp * unintrvn

. gen exp_10  = exp * wardur 

.  
. * interactions between trnsfcap and all other covariates
.  
. gen tran_1  = trnsfcap * treaty

. gen tran_2  = trnsfcap * factnum

. gen tran_3  = trnsfcap * factnum2

. gen tran_4  = trnsfcap * exp

. gen tran_5  = trnsfcap * logcost

. gen tran_6  = trnsfcap * develop

. gen tran_7  = trnsfcap * wartype

. gen tran_8  = trnsfcap * decade

. gen tran_9  = trnsfcap * unintrvn

. gen tran_10  = trnsfcap * wardur 

. 
. * interactions between decade and all other covariates
.  
. gen dec_1  = decade * treaty

. gen dec_2  = decade * factnum

. gen dec_3  = decade * factnum2

. gen dec_4  = decade * exp

. gen dec_5  = decade * logcost

. gen dec_6  = decade * develop

. gen dec_7  = decade * wartype

. gen dec_8  = decade * trnsfcap 

. gen dec_9  = decade * unintrvn

. gen dec_10  = decade * wardur 

. 
. * interactions between factnum and all other covariates
.  
. gen fact_1  = factnum * treaty

. gen fact_2  = factnum * logcost

. * no interaction with factnum2
. gen fact_4  = factnum * trnsfcap

. gen fact_5  = factnum * develop

. gen fact_6  = factnum * exp

. gen fact_7  = factnum * wartype

. gen fact_8  = factnum * decade

. gen fact_9  = factnum * unintrvn

. gen fact_10  = factnum * wardur 

. 
. 
. 
. * Model including wartype interactions
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn typetre
>  typefac typefac2 typetra typedev typeexp typedur typedec typeun1 typecost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      74.55
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -39.880988                Pseudo R2       =     0.4963

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.543758    4.98005    -0.51   0.609    -12.30448    7.216962
     logcost |  -.9161922   .5426285    -1.69   0.091    -1.979724      .14734
      wardur |   .0154774   .0127197     1.22   0.224    -.0094527    .0404075
     factnum |  -1.702182   2.662126    -0.64   0.523    -6.919854    3.515489
    factnum2 |   .1495605   .2673186     0.56   0.576    -.3743743    .6734953
    trnsfcap |    .032166   .0153338     2.10   0.036     .0021123    .0622197
     develop |   .0016444   .0010371     1.59   0.113    -.0003884    .0036771
         exp |  -.7684581   3.177037    -0.24   0.809    -6.995336     5.45842
      decade |  -.4460325   .3776486    -1.18   0.238     -1.18621    .2941453
      treaty |   2.635781   2.204708     1.20   0.232    -1.685368    6.956929
    unintrvn |   3.172576   1.401551     2.26   0.024     .4255863    5.919566
     typetre |   .1043365   2.564257     0.04   0.968    -4.921514    5.130187
     typefac |  -.2302627   2.833175    -0.08   0.935    -5.783183    5.322658
    typefac2 |  -.0312007   .2784769    -0.11   0.911    -.5770053     .514604
     typetra |  -.0282603   .0152837    -1.85   0.064    -.0582158    .0016952
     typedev |   -.001295    .001213    -1.07   0.286    -.0036725    .0010824
     typeexp |  -7.901738   5.659825    -1.40   0.163    -18.99479    3.191316
     typedur |  -.0078667   .0117865    -0.67   0.504    -.0309679    .0152345
     typedec |   .2774802   .4455956     0.62   0.533    -.5958713    1.150832
     typeun1 |  -2.015885   1.688832    -1.19   0.233    -5.325934    1.294165
    typecost |   .3648392   .6356091     0.57   0.566    -.8809318     1.61061
       _cons |   11.46168   3.336171     3.44   0.001      4.92291    18.00046
------------------------------------------------------------------------------

. test typetre typefac typefac2 typetra typedev typeexp typedur typedec typeun1 typecost

 ( 1)  typetre = 0
 ( 2)  typefac = 0
 ( 3)  typefac2 = 0
 ( 4)  typetra = 0
 ( 5)  typedev = 0
 ( 6)  typeexp = 0
 ( 7)  typedur = 0
 ( 8)  typedec = 0
 ( 9)  typeun1 = 0
 (10)  typecost = 0

           chi2( 10) =   10.78
         Prob > chi2 =    0.3750

. 
. * Model including treaty interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn tretype
>  trefac trefac2 tretra tredev treexp tredur tredec treun1 trecost, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -47.381221
Iteration 2:   log pseudo-likelihood = -41.945335
Iteration 3:   log pseudo-likelihood = -40.704533
Iteration 4:   log pseudo-likelihood = -40.560032
Iteration 5:   log pseudo-likelihood = -40.540271
Iteration 6:   log pseudo-likelihood = -40.539921
Iteration 7:   log pseudo-likelihood = -40.539921

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      58.52
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -40.539921                Pseudo R2       =     0.4880

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.225581   .8697375    -2.56   0.011    -3.930235   -.5209267
     logcost |  -.5344114   .2211727    -2.42   0.016     -.967902   -.1009209
      wardur |   .0126937   .0079985     1.59   0.113    -.0029831    .0283705
     factnum |  -3.090911   1.862852    -1.66   0.097    -6.742033    .5602121
    factnum2 |   .2562213   .2206002     1.16   0.245    -.1761473    .6885898
    trnsfcap |   .0062747   .0039569     1.59   0.113    -.0014807    .0140301
     develop |   .0011103   .0005772     1.92   0.054    -.0000209    .0022416
         exp |  -3.773393   5.088175    -0.74   0.458    -13.74603    6.199248
      decade |  -.3033613    .229925    -1.32   0.187     -.754006    .1472833
      treaty |   8.505266   8.753313     0.97   0.331    -8.650913    25.66144
    unintrvn |   1.403725   1.506694     0.93   0.352     -1.54934     4.35679
     tretype |   .5500494    1.67829     0.33   0.743    -2.739338    3.839437
      trefac |   3.814074   2.655854     1.44   0.151    -1.391304    9.019452
     trefac2 |  -.3959497   .3009285    -1.32   0.188    -.9857587    .1938592
      tretra |  -.0036393   .0049088    -0.74   0.458    -.0132604    .0059818
      tredev |   -.001282   .0009007    -1.42   0.155    -.0030474    .0004834
      treexp |  -8.950934   7.377943    -1.21   0.225    -23.41144    5.509568
      tredur |  -.0072861   .0086066    -0.85   0.397    -.0241548    .0095826
      tredec |     -.3395   .4152967    -0.82   0.414    -1.153467    .4744666
      treun1 |   2.548063   1.867486     1.36   0.172    -1.112143    6.208269
     trecost |   -.847007   .4853926    -1.74   0.081    -1.798359     .104345
       _cons |   12.09514     4.7598     2.54   0.011     2.766101    21.42417
------------------------------------------------------------------------------

. test tretype trefac trefac2 tretra tredev treexp tredur tredec trecost treun1 

 ( 1)  tretype = 0
 ( 2)  trefac = 0
 ( 3)  trefac2 = 0
 ( 4)  tretra = 0
 ( 5)  tredev = 0
 ( 6)  treexp = 0
 ( 7)  tredur = 0
 ( 8)  tredec = 0
 ( 9)  trecost = 0
 (10)  treun1 = 0

           chi2( 10) =   16.77
         Prob > chi2 =    0.0796

. 
. * Model including factnum interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn fact_1 
> fact_2 fact_4 fact_5 fact_6 fact_7 fact_8 fact_9 fact_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -46.476464
Iteration 2:   log pseudo-likelihood = -38.917673
Iteration 3:   log pseudo-likelihood = -35.412426
Iteration 4:   log pseudo-likelihood = -33.795017
Iteration 5:   log pseudo-likelihood = -33.152189
Iteration 6:   log pseudo-likelihood = -32.907491
Iteration 7:   log pseudo-likelihood = -32.882845
Iteration 8:   log pseudo-likelihood = -32.882553
Iteration 9:   log pseudo-likelihood = -32.882553

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =      45.70
                                                  Prob > chi2     =     0.0009
Log pseudo-likelihood = -32.882553                Pseudo R2       =     0.5847

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   2.902487   3.302746     0.88   0.380    -3.570776     9.37575
     logcost |  -.2819032   .8417507    -0.33   0.738    -1.931704    1.367898
      wardur |  -.0381123    .024686    -1.54   0.123     -.086496    .0102713
     factnum |    6.40135   6.621641     0.97   0.334    -6.576829    19.37953
    factnum2 |   -.874193   .5870386    -1.49   0.136    -2.024767    .2763815
    trnsfcap |   .0404036   .0206597     1.96   0.051    -.0000886    .0808958
     develop |   .0048024   .0018748     2.56   0.010     .0011278    .0084771
         exp |   10.14896   12.06403     0.84   0.400    -13.49611    33.79403
      decade |   1.763253   1.442474     1.22   0.222    -1.063943    4.590449
      treaty |  -6.065028   5.044905    -1.20   0.229    -15.95286    3.822804
    unintrvn |  -3.998014   5.126277    -0.78   0.435    -14.04533    6.049304
      fact_1 |   3.785069   2.481136     1.53   0.127    -1.077868    8.648007
      fact_2 |  -.1238268   .4069754    -0.30   0.761    -.9214839    .6738303
      fact_4 |  -.0059211   .0038926    -1.52   0.128    -.0135506    .0017083
      fact_5 |  -.0014927   .0006824    -2.19   0.029    -.0028302   -.0001552
      fact_6 |  -6.632362   4.807925    -1.38   0.168    -16.05572    2.790998
      fact_7 |  -1.629439   1.405685    -1.16   0.246    -4.384531    1.125652
      fact_8 |  -.9363034   .6384205    -1.47   0.142    -2.187585    .3149779
      fact_9 |    2.51903   2.320568     1.09   0.278    -2.029199    7.067259
     fact_10 |   .0202576   .0114155     1.77   0.076    -.0021164    .0426316
       _cons |  -4.149292   11.66738    -0.36   0.722    -27.01695    18.71836
------------------------------------------------------------------------------

note: 9 failures and 1 success completely determined.

. test fact_1 fact_2 fact_4 fact_5 fact_6 fact_7 fact_8 fact_9 fact_10

 ( 1)  fact_1 = 0
 ( 2)  fact_2 = 0
 ( 3)  fact_4 = 0
 ( 4)  fact_5 = 0
 ( 5)  fact_6 = 0
 ( 6)  fact_7 = 0
 ( 7)  fact_8 = 0
 ( 8)  fact_9 = 0
 ( 9)  fact_10 = 0

           chi2(  9) =   12.13
         Prob > chi2 =    0.2062

. 
. * Model including decade interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn dec_1 d
> ec_2 dec_3 dec_4 dec_5 dec_6 dec_7 dec_8 dec_9 dec_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -47.114022
Iteration 2:   log pseudo-likelihood = -42.538137
Iteration 3:   log pseudo-likelihood = -41.538962
Iteration 4:   log pseudo-likelihood = -41.466392
Iteration 5:   log pseudo-likelihood = -41.465734
Iteration 6:   log pseudo-likelihood = -41.465734

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      52.86
                                                  Prob > chi2     =     0.0001
Log pseudo-likelihood = -41.465734                Pseudo R2       =     0.4763

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.732101   1.382611    -1.98   0.048    -5.441968   -.0222337
     logcost |    -.97513   .5364173    -1.82   0.069    -2.026488    .0762285
      wardur |   .0403079   .0222162     1.81   0.070     -.003235    .0838508
     factnum |   .0328804     2.3783     0.01   0.989    -4.628502    4.694262
    factnum2 |  -.2681564   .3146385    -0.85   0.394    -.8848366    .3485237
    trnsfcap |    .017408   .0075377     2.31   0.021     .0026345    .0321816
     develop |  -.0006219    .001116    -0.56   0.577    -.0028092    .0015653
         exp |    2.03335   8.029501     0.25   0.800    -13.70418    17.77088
      decade |   .1994559   1.532627     0.13   0.896    -2.804437    3.203349
      treaty |   1.744524   2.121053     0.82   0.411    -2.412664    5.901712
    unintrvn |   6.064615   2.724106     2.23   0.026     .7254645    11.40377
       dec_1 |   .1715918    .540566     0.32   0.751     -.887898    1.231082
       dec_2 |   -.649917   .8023245    -0.81   0.418    -2.222444      .92261
       dec_3 |   .1285934   .1088771     1.18   0.238    -.0848018    .3419885
       dec_4 |  -2.488369   1.655527    -1.50   0.133    -5.733143    .7564053
       dec_5 |    .100485   .1346071     0.75   0.455    -.1633402    .3643101
       dec_6 |   .0001611    .000313     0.51   0.607    -.0004524    .0007747
       dec_7 |   .2537386    .343175     0.74   0.460    -.4188721    .9263493
       dec_8 |  -.0043069   .0022708    -1.90   0.058    -.0087575    .0001437
       dec_9 |  -1.012451   .6335617    -1.60   0.110    -2.254209    .2293074
      dec_10 |  -.0096969   .0056305    -1.72   0.085    -.0207326    .0013387
       _cons |   10.42244    6.66035     1.56   0.118    -2.631605    23.47649
------------------------------------------------------------------------------

. test dec_1 dec_2 dec_3 dec_4 dec_5 dec_6 dec_7 dec_8 dec_9 dec_10

 ( 1)  dec_1 = 0
 ( 2)  dec_2 = 0
 ( 3)  dec_3 = 0
 ( 4)  dec_4 = 0
 ( 5)  dec_5 = 0
 ( 6)  dec_6 = 0
 ( 7)  dec_7 = 0
 ( 8)  dec_8 = 0
 ( 9)  dec_9 = 0
 (10)  dec_10 = 0

           chi2( 10) =    8.05
         Prob > chi2 =    0.6240

. 
. * Model including exp interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn exp_1 e
> xp_2 exp_3 exp_4 exp_5 exp_6 exp_7 exp_8 exp_9 exp_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -49.127683
Iteration 2:   log pseudo-likelihood = -43.222733
Iteration 3:   log pseudo-likelihood = -40.102156
Iteration 4:   log pseudo-likelihood = -39.297478
Iteration 5:   log pseudo-likelihood = -39.233783
Iteration 6:   log pseudo-likelihood = -39.233281
Iteration 7:   log pseudo-likelihood = -39.233281

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      74.73
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -39.233281                Pseudo R2       =     0.5045

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -.6909128   1.289908    -0.54   0.592    -3.219086     1.83726
     logcost |   .2133613   .3115583     0.68   0.493    -.3972816    .8240043
      wardur |   .0029334   .0127754     0.23   0.818     -.022106    .0279728
     factnum |   2.457513   1.601205     1.53   0.125    -.6807917    5.595818
    factnum2 |  -.3028755   .1843765    -1.64   0.100    -.6642469    .0584959
    trnsfcap |   .0059271   .0066863     0.89   0.375    -.0071777     .019032
     develop |   .0004972   .0007778     0.64   0.523    -.0010274    .0020217
         exp |   188.6078   64.32592     2.93   0.003     62.53126    314.6842
      decade |   .2056162   .3599402     0.57   0.568    -.4998537    .9110861
      treaty |    2.29551   1.570244     1.46   0.144    -.7821111    5.373131
    unintrvn |  -1.801123   1.490251    -1.21   0.227    -4.721961    1.119715
       exp_1 |   1.567161   17.51364     0.09   0.929    -32.75895    35.89327
       exp_2 |  -50.18704   20.34274    -2.47   0.014    -90.05807     -10.316
       exp_3 |    4.92006   2.241239     2.20   0.028     .5273121    9.312809
       exp_4 |  -.0374781   .0793307    -0.47   0.637    -.1929634    .1180072
       exp_5 |  -8.296599   3.346154    -2.48   0.013    -14.85494   -1.738257
       exp_6 |   .0027519   .0083556     0.33   0.742    -.0136248    .0191286
       exp_7 |  -22.58431   15.61421    -1.45   0.148     -53.1876    8.018974
       exp_8 |  -3.480693   4.196712    -0.83   0.407     -11.7061    4.744712
       exp_9 |      46.88   20.77365     2.26   0.024      6.16439     87.5956
      exp_10 |   .0371259   .0994877     0.37   0.709    -.1578664    .2321181
       _cons |  -7.019536   5.531631    -1.27   0.204    -17.86133    3.822262
------------------------------------------------------------------------------

note: 2 failures and 0 successes completely determined.

. test exp_1 exp_2 exp_3 exp_4 exp_5 exp_6 exp_7 exp_8 exp_9 exp_10

 ( 1)  exp_1 = 0
 ( 2)  exp_2 = 0
 ( 3)  exp_3 = 0
 ( 4)  exp_4 = 0
 ( 5)  exp_5 = 0
 ( 6)  exp_6 = 0
 ( 7)  exp_7 = 0
 ( 8)  exp_8 = 0
 ( 9)  exp_9 = 0
 (10)  exp_10 = 0

           chi2( 10) =   10.00
         Prob > chi2 =    0.4408

. 
. * Model including develop interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn dev_1 d
> ev_2 dev_3 dev_4 dev_5 dev_6 dev_7 dev_8 dev_9 dev_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -47.348777
Iteration 2:   log pseudo-likelihood = -41.874455
Iteration 3:   log pseudo-likelihood =  -39.73239
Iteration 4:   log pseudo-likelihood = -38.717072
Iteration 5:   log pseudo-likelihood = -38.323132
Iteration 6:   log pseudo-likelihood = -38.274929
Iteration 7:   log pseudo-likelihood = -38.274249
Iteration 8:   log pseudo-likelihood = -38.274249

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      70.60
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -38.274249                Pseudo R2       =     0.5166

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -.5675886   .7867813    -0.72   0.471    -2.109652    .9744744
     logcost |  -.1797839   .1779118    -1.01   0.312    -.5284846    .1689169
      wardur |   .0037775    .007397     0.51   0.610    -.0107203    .0182753
     factnum |  -.2545656   .7369396    -0.35   0.730    -1.698941    1.189809
    factnum2 |  -.0043606    .060252    -0.07   0.942    -.1224523    .1137312
    trnsfcap |   .0024422    .004298     0.57   0.570    -.0059818    .0108662
     develop |    .035291   .0156782     2.25   0.024     .0045623    .0660196
         exp |  -12.71412   5.128936    -2.48   0.013    -22.76665   -2.661588
      decade |  -.1624885   .2249748    -0.72   0.470    -.6034311    .2784541
      treaty |   1.800518   .9462594     1.90   0.057    -.0541159    3.655153
    unintrvn |   1.046644   .9713113     1.08   0.281    -.8570917    2.950379
       dev_1 |   .0033891    .002847     1.19   0.234    -.0021909    .0089691
       dev_2 |   -.006727   .0038254    -1.76   0.079    -.0142247    .0007707
       dev_3 |    .000584   .0003724     1.57   0.117    -.0001459     .001314
       dev_4 |   8.38e-06   6.43e-06     1.30   0.193    -4.23e-06     .000021
       dev_5 |  -.0017266   .0009078    -1.90   0.057    -.0035058    .0000527
       dev_6 |   .0159117   .0190227     0.84   0.403    -.0213722    .0531956
       dev_7 |  -.0081108   .0028512    -2.84   0.004    -.0136991   -.0025226
       dev_8 |   .0001125   .0005188     0.22   0.828    -.0009042    .0011293
       dev_9 |   .0041163   .0024117     1.71   0.088    -.0006106    .0088431
      dev_10 |   5.25e-06   .0000121     0.43   0.664    -.0000184    .0000289
       _cons |   2.935425   2.529284     1.16   0.246    -2.021879     7.89273
------------------------------------------------------------------------------

note: 2 failures and 3 successes completely determined.

. test dev_1 dev_2 dev_3 dev_4 dev_5 dev_6 dev_7 dev_8 dev_9 dev_10

 ( 1)  dev_1 = 0
 ( 2)  dev_2 = 0
 ( 3)  dev_3 = 0
 ( 4)  dev_4 = 0
 ( 5)  dev_5 = 0
 ( 6)  dev_6 = 0
 ( 7)  dev_7 = 0
 ( 8)  dev_8 = 0
 ( 9)  dev_9 = 0
 (10)  dev_10 = 0

           chi2( 10) =   15.24
         Prob > chi2 =    0.1234

. 
. * Model with logcost interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn cost_1 
> cost_2 cost_3 cost_4 cost_5 cost_6 cost_7 cost_8 cost_9 cost_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -44.063892
Iteration 2:   log pseudo-likelihood = -37.305176
Iteration 3:   log pseudo-likelihood = -34.624382
Iteration 4:   log pseudo-likelihood = -34.046541
Iteration 5:   log pseudo-likelihood = -33.995896
Iteration 6:   log pseudo-likelihood = -33.995433
Iteration 7:   log pseudo-likelihood = -33.995433

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      56.59
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -33.995433                Pseudo R2       =     0.5706

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -4.007538   3.168356    -1.26   0.206     -10.2174    2.202326
     logcost |  -5.798863   1.826972    -3.17   0.002    -9.379662   -2.218064
      wardur |   .0548272   .0338186     1.62   0.105     -.011456    .1211104
     factnum |   -50.8783   17.62667    -2.89   0.004    -85.42594   -16.33066
    factnum2 |   6.870291   2.524614     2.72   0.007     1.922139    11.81844
    trnsfcap |  -.0049852   .0232075    -0.21   0.830    -.0504711    .0405007
     develop |   .0113591   .0051148     2.22   0.026     .0013342    .0213839
         exp |   70.24995   39.23829     1.79   0.073    -6.655678    147.1556
      decade |  -.7737324   .9453909    -0.82   0.413    -2.626665      1.0792
      treaty |  -.3407935   7.947156    -0.04   0.966    -15.91693    15.23535
    unintrvn |    5.19413   7.101236     0.73   0.465    -8.724036     19.1123
      cost_1 |   .2320783   .6046968     0.38   0.701    -.9531056    1.417262
      cost_2 |   3.893788   1.367642     2.85   0.004     1.213259    6.574316
      cost_3 |  -.5328627   .1965379    -2.71   0.007    -.9180699   -.1476556
      cost_4 |   .0010391   .0016302     0.64   0.524    -.0021561    .0042342
      cost_5 |   -.000861    .000384    -2.24   0.025    -.0016136   -.0001084
      cost_6 |  -6.954962   3.777088    -1.84   0.066    -14.35792    .4479939
      cost_7 |   .0874856   .3006304     0.29   0.771     -.501739    .6767103
      cost_8 |   .0618231    .089338     0.69   0.489    -.1132762    .2369224
      cost_9 |  -.2419507   .5503405    -0.44   0.660    -1.320598    .8366969
     cost_10 |  -.0032852   .0024075    -1.36   0.172    -.0080038    .0014334
       _cons |   76.17247   23.87256     3.19   0.001     29.38311    122.9618
------------------------------------------------------------------------------

note: 0 failures and 1 success completely determined.

. test cost_1 cost_2 cost_3 cost_4 cost_5 cost_6 cost_7 cost_8 cost_9 cost_10

 ( 1)  cost_1 = 0
 ( 2)  cost_2 = 0
 ( 3)  cost_3 = 0
 ( 4)  cost_4 = 0
 ( 5)  cost_5 = 0
 ( 6)  cost_6 = 0
 ( 7)  cost_7 = 0
 ( 8)  cost_8 = 0
 ( 9)  cost_9 = 0
 (10)  cost_10 = 0

           chi2( 10) =   18.65
         Prob > chi2 =    0.0450

. 
. * Model with trnsfcap interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn tran_1 
> tran_2 tran_3 tran_4 tran_5 tran_6 tran_7 tran_8 tran_9 tran_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood =  -46.08797
Iteration 2:   log pseudo-likelihood =  -40.86495
Iteration 3:   log pseudo-likelihood = -39.045337
Iteration 4:   log pseudo-likelihood = -37.714316
Iteration 5:   log pseudo-likelihood = -37.184712
Iteration 6:   log pseudo-likelihood = -36.985886
Iteration 7:   log pseudo-likelihood = -36.955475
Iteration 8:   log pseudo-likelihood =  -36.93625
Iteration 9:   log pseudo-likelihood = -36.936175

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      77.82
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -36.936175                Pseudo R2       =     0.5335

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -.2882652   .8259283    -0.35   0.727    -1.907055    1.330524
     logcost |  -.6784838   .1462524    -4.64   0.000    -.9651332   -.3918344
      wardur |   .0093639    .007238     1.29   0.196    -.0048222    .0235501
     factnum |   .1893502    1.15686     0.16   0.870    -2.078054    2.456754
    factnum2 |   -.068973   .1302436    -0.53   0.596    -.3242457    .1862998
    trnsfcap |   .3299737   .1217571     2.71   0.007     .0913341    .5686133
     develop |   .0006598   .0005619     1.17   0.240    -.0004415    .0017611
         exp |  -5.288844     4.8951    -1.08   0.280    -14.88306    4.305376
      decade |  -.0979776   .2548316    -0.38   0.701    -.5974384    .4014832
      treaty |   1.794368   1.141287     1.57   0.116    -.4425122    4.031249
    unintrvn |   1.090504    .985564     1.11   0.269    -.8411661    3.022174
      tran_1 |   .0839535   .0513442     1.64   0.102    -.0166792    .1845863
      tran_2 |  -.1356188    .058009    -2.34   0.019    -.2493144   -.0219232
      tran_3 |   .0107612   .0056959     1.89   0.059    -.0004027     .021925
      tran_4 |  -.1482334   .2948699    -0.50   0.615    -.7261678     .429701
      tran_5 |   .0071262   .0110983     0.64   0.521     -.014626    .0288784
      tran_6 |   9.76e-06   .0000238     0.41   0.681    -.0000368    .0000564
      tran_7 |   -.078873    .045674    -1.73   0.084    -.1683924    .0106463
      tran_8 |  -.0166927   .0100621    -1.66   0.097     -.036414    .0030286
      tran_9 |   .0545688   .0659732     0.83   0.408    -.0747364    .1838739
     tran_10 |   .0000139   .0001734     0.08   0.936    -.0003259    .0003537
       _cons |   5.918347   2.508518     2.36   0.018     1.001742    10.83495
------------------------------------------------------------------------------

note: 2 failures and 3 successes completely determined.

. test tran_1 tran_2 tran_3 tran_4 tran_5 tran_6 tran_7 tran_8 tran_9 tran_10

 ( 1)  tran_1 = 0
 ( 2)  tran_2 = 0
 ( 3)  tran_3 = 0
 ( 4)  tran_4 = 0
 ( 5)  tran_5 = 0
 ( 6)  tran_6 = 0
 ( 7)  tran_7 = 0
 ( 8)  tran_8 = 0
 ( 9)  tran_9 = 0
 (10)  tran_10 = 0

           chi2( 10) =   24.33
         Prob > chi2 =    0.0068

. 
. * Model with war duration interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn dur_1 d
> ur_2 dur_3 dur_4 dur_5 dur_6 dur_7 dur_8 dur_9 dur_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -47.396044
Iteration 2:   log pseudo-likelihood = -41.013563
Iteration 3:   log pseudo-likelihood = -39.039977
Iteration 4:   log pseudo-likelihood = -38.615876
Iteration 5:   log pseudo-likelihood = -38.581377
Iteration 6:   log pseudo-likelihood = -38.580839
Iteration 7:   log pseudo-likelihood = -38.580839

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      45.22
                                                  Prob > chi2     =     0.0016
Log pseudo-likelihood = -38.580839                Pseudo R2       =     0.5127

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.239167    .952665    -2.35   0.019    -4.106356   -.3719781
     logcost |  -.4221542   .2815389    -1.50   0.134    -.9739602    .1296518
      wardur |   .0545304   .0538001     1.01   0.311    -.0509158    .1599767
     factnum |  -3.585532    2.34185    -1.53   0.126    -8.175473     1.00441
    factnum2 |   .1978648   .2776753     0.71   0.476    -.3463687    .7420983
    trnsfcap |   .0077015   .0027289     2.82   0.005      .002353    .0130501
     develop |   .0002882   .0005801     0.50   0.619    -.0008487    .0014252
         exp |  -1.714688   5.301935    -0.32   0.746    -12.10629    8.676913
      decade |  -.0503933   .2665353    -0.19   0.850    -.5727929    .4720062
      treaty |   4.131853   1.649419     2.51   0.012     .8990512    7.364654
    unintrvn |   .9175729   1.667189     0.55   0.582    -2.350058    4.185204
       dur_1 |  -.0096878   .0122201    -0.79   0.428    -.0336388    .0142632
       dur_2 |   .0242516   .0216872     1.12   0.263    -.0182545    .0667577
       dur_3 |  -.0017721   .0022154    -0.80   0.424    -.0061143    .0025701
       dur_4 |  -.0000589   .0000436    -1.35   0.177    -.0001444    .0000266
       dur_5 |   6.24e-06   8.48e-06     0.74   0.461    -.0000104    .0000229
       dur_6 |  -.1393018   .0879887    -1.58   0.113    -.3117566    .0331529
       dur_7 |   .0060293   .0119533     0.50   0.614    -.0173988    .0294574
       dur_8 |  -.0127476   .0052676    -2.42   0.016    -.0230719   -.0024234
       dur_9 |   .0252112    .014113     1.79   0.074    -.0024499    .0528722
      dur_10 |  -.0040632   .0029528    -1.38   0.169    -.0098506    .0017242
       _cons |   11.76442   5.792286     2.03   0.042     .4117475    23.11709
------------------------------------------------------------------------------

. test dur_1 dur_2 dur_3 dur_4 dur_5 dur_6 dur_7 dur_8 dur_9 dur_10

 ( 1)  dur_1 = 0
 ( 2)  dur_2 = 0
 ( 3)  dur_3 = 0
 ( 4)  dur_4 = 0
 ( 5)  dur_5 = 0
 ( 6)  dur_6 = 0
 ( 7)  dur_7 = 0
 ( 8)  dur_8 = 0
 ( 9)  dur_9 = 0
 (10)  dur_10 = 0

           chi2( 10) =   12.94
         Prob > chi2 =    0.2272

. 
. * Same model, dropping decade because wars that started later will have lower war duration
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp treaty unintrvn dur_1 dur_2 du
> r_3 dur_4 dur_5 dur_6 dur_7 dur_9 dur_10, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -49.056174
Iteration 2:   log pseudo-likelihood = -44.224038
Iteration 3:   log pseudo-likelihood = -43.216816
Iteration 4:   log pseudo-likelihood = -43.119285
Iteration 5:   log pseudo-likelihood = -43.115765
Iteration 6:   log pseudo-likelihood =  -43.11576

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(19)   =      53.76
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood =  -43.11576                Pseudo R2       =     0.4554

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.327203   .9665766    -2.41   0.016    -4.221658   -.4327475
     logcost |  -.4176154   .2283494    -1.83   0.067    -.8651721    .0299412
      wardur |  -.0067121   .0483774    -0.14   0.890    -.1015301    .0881059
     factnum |  -3.545732   2.287119    -1.55   0.121    -8.028404     .936939
    factnum2 |   .2565577   .2849785     0.90   0.368    -.3019899    .8151053
    trnsfcap |   .0063588   .0019876     3.20   0.001     .0024631    .0102545
     develop |   .0006711   .0006252     1.07   0.283    -.0005542    .0018964
         exp |  -6.170129   4.694903    -1.31   0.189    -15.37197    3.031711
      treaty |   2.476552   1.203016     2.06   0.040     .1186839    4.834421
    unintrvn |   1.292622   1.420084     0.91   0.363     -1.49069    4.075935
       dur_1 |  -.0041303   .0113623    -0.36   0.716    -.0263999    .0181393
       dur_2 |   .0213038   .0183812     1.16   0.246    -.0147227    .0573304
       dur_3 |  -.0018781   .0019317    -0.97   0.331    -.0056642    .0019081
       dur_4 |  -.0000159   .0000291    -0.55   0.584     -.000073    .0000411
       dur_5 |  -3.75e-08   6.70e-06    -0.01   0.996    -.0000132    .0000131
       dur_6 |   -.010268   .0646162    -0.16   0.874    -.1369134    .1163773
       dur_7 |   .0024811   .0117414     0.21   0.833    -.0205317    .0254939
       dur_9 |   .0068613   .0106357     0.65   0.519    -.0139844     .027707
      dur_10 |   -.002212   .0022353    -0.99   0.322    -.0065931     .002169
       _cons |   11.58268   5.331795     2.17   0.030     1.132549     22.0328
------------------------------------------------------------------------------

. test dur_1 dur_2 dur_3 dur_4 dur_5 dur_6 dur_7 dur_9 dur_10

 ( 1)  dur_1 = 0
 ( 2)  dur_2 = 0
 ( 3)  dur_3 = 0
 ( 4)  dur_4 = 0
 ( 5)  dur_5 = 0
 ( 6)  dur_6 = 0
 ( 7)  dur_7 = 0
 ( 8)  dur_9 = 0
 ( 9)  dur_10 = 0

           chi2(  9) =    9.93
         Prob > chi2 =    0.3563

. 
. 
. ** Interactions with high/low versions of non-binary variables
. 
. * Model with high/low factnum interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn facttyp
> e facttre facttra factdev factexp factdur factdec factun1 factcost, cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -47.507237
Iteration 2:   log pseudo-likelihood =  -41.74538
Iteration 3:   log pseudo-likelihood = -39.606997
Iteration 4:   log pseudo-likelihood = -39.008304
Iteration 5:   log pseudo-likelihood = -38.927332
Iteration 6:   log pseudo-likelihood = -38.924639
Iteration 7:   log pseudo-likelihood = -38.924633

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =      77.82
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -38.924633                Pseudo R2       =     0.5084

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -6.713935   2.055016    -3.27   0.001    -10.74169   -2.686178
     logcost |  -.8882017   .2187697    -4.06   0.000    -1.316982    -.459421
      wardur |    .018651   .0089259     2.09   0.037     .0011566    .0361455
     factnum |  -3.106568    1.13618    -2.73   0.006    -5.333439   -.8796962
    factnum2 |   .2403048   .1037252     2.32   0.021     .0370071    .4436024
    trnsfcap |   .0109405   .0038182     2.87   0.004      .003457    .0184241
     develop |    .000536   .0005786     0.93   0.354     -.000598    .0016701
         exp |  -10.40146   5.827089    -1.79   0.074    -21.82235    1.019425
      decade |   .8875272   .6435993     1.38   0.168    -.3739042    2.148959
      treaty |   .0841617   1.903271     0.04   0.965    -3.646181    3.814505
    unintrvn |   4.608154    1.78959     2.57   0.010     1.100622    8.115685
    facttype |   2.193297   .7923765     2.77   0.006     .6402676    3.746327
     facttre |   1.002939   .7675472     1.31   0.191    -.5014256    2.507304
     facttra |   .0010331   .0027642     0.37   0.709    -.0043845    .0064508
     factdev |   .0000497   .0003592     0.14   0.890    -.0006543    .0007537
     factexp |   1.849257   2.816723     0.66   0.511    -3.671419    7.369933
     factdur |  -.0052665   .0040163    -1.31   0.190    -.0131383    .0026053
     factdec |  -.5385095   .2565089    -2.10   0.036    -1.041258   -.0357613
     factun1 |  -1.038478   .6614664    -1.57   0.116    -2.334928    .2579726
    factcost |   .1296939   .1067548     1.21   0.224    -.0795416    .3389295
       _cons |    13.1775   3.056431     4.31   0.000     7.187007      19.168
------------------------------------------------------------------------------

. test facttype facttre facttra factdev factexp factdur factdec factun1 factcost

 ( 1)  facttype = 0
 ( 2)  facttre = 0
 ( 3)  facttra = 0
 ( 4)  factdev = 0
 ( 5)  factexp = 0
 ( 6)  factdur = 0
 ( 7)  factdec = 0
 ( 8)  factun1 = 0
 ( 9)  factcost = 0

           chi2(  9) =   21.65
         Prob > chi2 =    0.0101

. 
. * Model with high/low decade interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn dectype
>  decfac decfac2 dectre dectra decdev decexp decdur decun1 deccost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      42.15
                                                  Prob > chi2     =     0.0040
Log pseudo-likelihood = -42.536834                Pseudo R2       =     0.4627

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.796814   .7916434    -2.27   0.023    -3.348407    -.245222
     logcost |  -.6216346   .2610754    -2.38   0.017    -1.133333   -.1099362
      wardur |   .0143654   .0097095     1.48   0.139    -.0046648    .0333956
     factnum |  -2.201367   .8261152    -2.66   0.008    -3.820523   -.5822107
    factnum2 |   .1507837   .0666768     2.26   0.024     .0200997    .2814678
    trnsfcap |   .0066142   .0026878     2.46   0.014     .0013461    .0118822
     develop |    .000092   .0006085     0.15   0.880    -.0011006    .0012846
         exp |  -1.087532   5.770716    -0.19   0.851    -12.39793    10.22286
      decade |  -.6392287   .4846106    -1.32   0.187    -1.589048    .3105906
      treaty |   .8399132   1.253146     0.67   0.503    -1.616209    3.296035
    unintrvn |   3.299756   1.634903     2.02   0.044     .0954058    6.504107
     dectype |  -.2835264   1.005643    -0.28   0.778    -2.254551    1.687498
      decfac |   .9395872   1.265558     0.74   0.458    -1.540861    3.420035
     decfac2 |  -.0747539   .1452335    -0.51   0.607    -.3594064    .2098986
      dectre |   2.234356   1.590857     1.40   0.160    -.8836672    5.352378
      dectra |  -.0073469   .0041533    -1.77   0.077    -.0154872    .0007934
      decdev |   .0002745   .0008719     0.31   0.753    -.0014344    .0019835
      decexp |   -9.25287   6.845125    -1.35   0.176    -22.66907    4.163329
      decdur |  -.0110834    .012816    -0.86   0.387    -.0362024    .0140355
      decun1 |  -2.305835   1.750674    -1.32   0.188    -5.737093    1.125422
     deccost |    .112246   .2914477     0.39   0.700     -.458981    .6834731
       _cons |   11.65756   4.177148     2.79   0.005     3.470497    19.84462
------------------------------------------------------------------------------

. test dectype decfac decfac2 dectre dectra decdev decexp decdur decun1 deccost

 ( 1)  dectype = 0
 ( 2)  decfac = 0
 ( 3)  decfac2 = 0
 ( 4)  dectre = 0
 ( 5)  dectra = 0
 ( 6)  decdev = 0
 ( 7)  decexp = 0
 ( 8)  decdur = 0
 ( 9)  decun1 = 0
 (10)  deccost = 0

           chi2( 10) =    5.88
         Prob > chi2 =    0.8248

. 
. * Model with high/low exp interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn exptype
>  expfac expfac2 exptre exptra expdev expdur expdec expun1 expcost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      42.86
                                                  Prob > chi2     =     0.0033
Log pseudo-likelihood = -42.955562                Pseudo R2       =     0.4574

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.922524   .9244469    -2.08   0.038    -3.734407   -.1106415
     logcost |  -.5537388     .23758    -2.33   0.020    -1.019387   -.0880905
      wardur |   .0100745   .0086391     1.17   0.244    -.0068578    .0270069
     factnum |  -1.625808   .8071919    -2.01   0.044    -3.207875   -.0437413
    factnum2 |   .1027142   .0608213     1.69   0.091    -.0164934    .2219217
    trnsfcap |   .0038336   .0020579     1.86   0.062    -.0001998     .007867
     develop |   .0008894   .0004983     1.78   0.074    -.0000874    .0018661
         exp |  -11.63985   6.460626    -1.80   0.072    -24.30245    1.022741
      decade |    -.45359   .1999602    -2.27   0.023    -.8455048   -.0616751
      treaty |   1.848572   .9618606     1.92   0.055    -.0366401    3.733784
    unintrvn |   1.936181   1.054116     1.84   0.066    -.1298486    4.002211
     exptype |  -.6009228   1.337386    -0.45   0.653    -3.222151    2.020305
      expfac |   .6514288   1.716546     0.38   0.704     -2.71294    4.015798
     expfac2 |  -.0764746   .2425557    -0.32   0.753    -.5518752    .3989259
      exptre |   1.256046   1.349887     0.93   0.352    -1.389683    3.901776
      exptra |   .0147003   .0126847     1.16   0.246    -.0101612    .0395619
      expdev |   -.000144   .0008155    -0.18   0.860    -.0017424    .0014544
      expdur |  -.0023233   .0102712    -0.23   0.821    -.0224546    .0178079
      expdec |   .6784331   .5400966     1.26   0.209    -.3801368    1.737003
      expun1 |  -.8814659   1.320816    -0.67   0.505    -3.470219    1.707287
     expcost |  -.1973177   .2966706    -0.67   0.506    -.7787815    .3841461
       _cons |   10.62914   3.331869     3.19   0.001     4.098797    17.15948
------------------------------------------------------------------------------

. test exptype expfac expfac2 exptre exptra expdev expdur expdec expun1 expcost

 ( 1)  exptype = 0
 ( 2)  expfac = 0
 ( 3)  expfac2 = 0
 ( 4)  exptre = 0
 ( 5)  exptra = 0
 ( 6)  expdev = 0
 ( 7)  expdur = 0
 ( 8)  expdec = 0
 ( 9)  expun1 = 0
 (10)  expcost = 0

           chi2( 10) =    3.54
         Prob > chi2 =    0.9656

. 
. * Model with high/low develop interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn devtype
>  devfac devfac2 devtre devtra devexp devdur devdec devun1 devcost , cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      48.77
                                                  Prob > chi2     =     0.0005
Log pseudo-likelihood = -43.241347                Pseudo R2       =     0.4538

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.930109   .8449194    -2.28   0.022     -3.58612   -.2740971
     logcost |  -.5234025   .1802307    -2.90   0.004    -.8766481   -.1701569
      wardur |   .0056511   .0080151     0.71   0.481    -.0100582    .0213605
     factnum |  -1.394401   .7346415    -1.90   0.058    -2.834272    .0454697
    factnum2 |   .0913134   .0593778     1.54   0.124     -.025065    .2076918
    trnsfcap |   .0036234   .0033488     1.08   0.279    -.0029401    .0101869
     develop |   .0009129   .0007971     1.15   0.252    -.0006494    .0024752
         exp |  -9.419473   4.173553    -2.26   0.024    -17.59949    -1.23946
      decade |  -.3118548    .226857    -1.37   0.169    -.7564864    .1327768
      treaty |   2.455501   1.093196     2.25   0.025     .3128758    4.598126
    unintrvn |   1.980876   1.087084     1.82   0.068     -.149769    4.111521
     devtype |  -.2462969   1.366271    -0.18   0.857    -2.924139    2.431546
      devfac |   -.375087    2.08893    -0.18   0.857    -4.469315    3.719141
     devfac2 |  -.0238647   .2265549    -0.11   0.916    -.4679041    .4201748
      devtre |  -.5950656   1.795157    -0.33   0.740    -4.113508    2.923377
      devtra |   .0037209   .0038257     0.97   0.331    -.0037773    .0112192
      devexp |   12.71229    7.88892     1.61   0.107    -2.749704    28.17429
      devdur |    .008088   .0108737     0.74   0.457    -.0132239       .0294
      devdec |   .0883795   .3811303     0.23   0.817    -.6586222    .8353812
      devun1 |  -.6182612   1.247213    -0.50   0.620    -3.062753    1.826231
     devcost |  -.0757433   .3983306    -0.19   0.849     -.856457    .7049705
       _cons |   9.667702   2.504759     3.86   0.000     4.758464    14.57694
------------------------------------------------------------------------------

. test devtype devfac devfac2 devtre devtra devexp devdur devdec devun1 devcost 

 ( 1)  devtype = 0
 ( 2)  devfac = 0
 ( 3)  devfac2 = 0
 ( 4)  devtre = 0
 ( 5)  devtra = 0
 ( 6)  devexp = 0
 ( 7)  devdur = 0
 ( 8)  devdec = 0
 ( 9)  devun1 = 0
 (10)  devcost = 0

           chi2( 10) =   10.28
         Prob > chi2 =    0.4161

. 
. * Model with high/low logcost interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn costun1
>  costtype costfac costfac2 costtre costtra costdev costexp costdur costdec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(20)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -34.353488                Pseudo R2       =     0.5661

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.854144   1.356318    -2.10   0.035    -5.512479   -.1958088
     logcost |  -.6113108   .3136492    -1.95   0.051    -1.226052    .0034304
      wardur |   .0418748   .0137283     3.05   0.002     .0149678    .0687819
     factnum |  -3.723898   2.359251    -1.58   0.114    -8.347945    .9001496
    factnum2 |   .2093331   .4123683     0.51   0.612     -.598894     1.01756
    trnsfcap |   .0049462   .0044379     1.11   0.265    -.0037521    .0136444
     develop |   .0041492   .0016881     2.46   0.014     .0008405    .0074579
         exp |   2.330409   5.415814     0.43   0.667     -8.28439    12.94521
      decade |  -.1714098   .2674188    -0.64   0.522     -.695541    .3527213
      treaty |   22.10563   1.152702    19.18   0.000     19.84638    24.36489
    unintrvn |   .8231113   1.146553     0.72   0.473     -1.42409    3.070313
     costun1 |   1.345214   1.344307     1.00   0.317     -1.28958    3.980007
    costtype |  -.0681418    1.91339    -0.04   0.972    -3.818317    3.682033
     costfac |   2.648672   1.649147     1.61   0.108    -.5835961    5.880939
    costfac2 |  -.1710806   .3478533    -0.49   0.623    -.8528606    .5106995
     costtre |  -19.66302          .        .       .            .           .
     costtra |   .0003277   .0045425     0.07   0.942    -.0085753    .0092308
     costdev |  -.0042863   .0017586    -2.44   0.015    -.0077331   -.0008395
     costexp |  -16.57318   9.256287    -1.79   0.073    -34.71517    1.568807
     costdur |  -.0415752   .0153249    -2.71   0.007    -.0716114   -.0115391
     costdec |  -.0996802   .4061704    -0.25   0.806    -.8957596    .6963991
       _cons |   11.84717   5.579883     2.12   0.034     .9107973    22.78354
------------------------------------------------------------------------------

note: 0 failures and 9 successes completely determined.

. test costun1 costtype costfac costfac2 costtre costtra costdev costexp costdur costdec

 ( 1)  costun1 = 0
 ( 2)  costtype = 0
 ( 3)  costfac = 0
 ( 4)  costfac2 = 0
 ( 5)  costtre = 0
 ( 6)  costtra = 0
 ( 7)  costdev = 0
 ( 8)  costexp = 0
 ( 9)  costdur = 0
 (10)  costdec = 0
       Constraint 5 dropped

           chi2(  9) =   29.76
         Prob > chi2 =    0.0005

. 
. * Model with high/low trnsfcap interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn trantyp
> e tranfac tranfac2 trantre trandev tranexp trandur trandec tranun1 trancost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(18)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -37.718536                Pseudo R2       =     0.5236

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.233397   .6572501    -1.88   0.061    -2.521583    .0547894
     logcost |  -.5728776   .1756744    -3.26   0.001    -.9171932    -.228562
      wardur |   .0078529   .0061709     1.27   0.203    -.0042419    .0199477
     factnum |  -1.144981   .6818058    -1.68   0.093    -2.481296    .1913339
    factnum2 |   .0577152   .0553596     1.04   0.297    -.0507877     .166218
    trnsfcap |    .041051     .02119     1.94   0.053    -.0004806    .0825825
     develop |    .000749   .0004782     1.57   0.117    -.0001882    .0016862
         exp |  -4.929058   3.364188    -1.47   0.143    -11.52274    1.664628
      decade |   -.259418    .194206    -1.34   0.182    -.6400548    .1212188
      treaty |   2.578688   .8625581     2.99   0.003     .8881051    4.269271
    unintrvn |   1.113123   .7836165     1.42   0.155    -.4227375    2.648983
    trantype |  -1512.534   247.9618    -6.10   0.000     -1998.53   -1026.538
     tranfac |  -700.4141   122.6172    -5.71   0.000    -940.7394   -460.0887
    tranfac2 |    106.245   18.60226     5.71   0.000      69.7852    142.7047
     trantre |   74.48004   17.97333     4.14   0.000     39.25295    109.7071
     trandev |   .6882562   .1190857     5.78   0.000     .4548524    .9216599
     tranexp |   1072.766          .        .       .            .           .
     trandur |  -5.084051    .883708    -5.75   0.000    -6.816087   -3.352015
     trandec |   61.54511    11.4101     5.39   0.000     39.18173    83.90849
     tranun1 |   206.5561   31.23854     6.61   0.000     145.3297    267.7825
    trancost |   63.17032   12.43641     5.08   0.000      38.7954    87.54525
       _cons |   8.400052   2.272191     3.70   0.000     3.946639    12.85346
------------------------------------------------------------------------------

note: 6 failures and 5 successes completely determined.

. test trantype tranfac tranfac2 trantre trandev tranexp trandur trandec tranun1 trancost

 ( 1)  trantype = 0
 ( 2)  tranfac = 0
 ( 3)  tranfac2 = 0
 ( 4)  trantre = 0
 ( 5)  trandev = 0
 ( 6)  tranexp = 0
 ( 7)  trandur = 0
 ( 8)  trandec = 0
 ( 9)  tranun1 = 0
 (10)  trancost = 0
       Constraint 6 dropped

           chi2(  9) =47898.43
         Prob > chi2 =    0.0000

. 
. * Model with high/low war duration interactions
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty unintrvn durtype
>  durfac durfac2 durtre durtra durdev durexp durcost durdec durun1 , cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(21)   =      44.64
                                                  Prob > chi2     =     0.0019
Log pseudo-likelihood = -37.171496                Pseudo R2       =     0.5305

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.813664   .8744104    -2.07   0.038    -3.527477   -.0998515
     logcost |  -.5178365   .2507766    -2.06   0.039     -1.00935   -.0263234
      wardur |   .0080043   .0061257     1.31   0.191    -.0040018    .0200104
     factnum |  -3.449877   2.102191    -1.64   0.101    -7.570095    .6703417
    factnum2 |   .1506819   .2815514     0.54   0.593    -.4011487    .7025125
    trnsfcap |   .0093352   .0037212     2.51   0.012     .0020419    .0166286
     develop |   .0009128   .0006358     1.44   0.151    -.0003333     .002159
         exp |  -6.058056   4.257813    -1.42   0.155    -14.40322    2.287104
      decade |  -.3703081   .2501897    -1.48   0.139     -.860671    .1200547
      treaty |   3.460146   1.472248     2.35   0.019     .5745922    6.345699
    unintrvn |   2.047595   1.570214     1.30   0.192    -1.029967    5.125157
     durtype |   .2787623   1.289768     0.22   0.829    -2.249137    2.806662
      durfac |   3.424816   1.916288     1.79   0.074    -.3310388    7.180671
     durfac2 |  -.1734622   .2598044    -0.67   0.504    -.6826695    .3357451
      durtre |   -2.20426    1.80796    -1.22   0.223    -5.747796    1.339276
      durtra |   -.014958   .0053377    -2.80   0.005    -.0254197   -.0044964
      durdev |   .0006866   .0010456     0.66   0.511    -.0013628    .0027361
      durexp |  -9.469108   8.092203    -1.17   0.242    -25.32953    6.391318
     durcost |  -.2738314   .2693889    -1.02   0.309     -.801824    .2541612
      durdec |  -1.021409   .5388116    -1.90   0.058     -2.07746    .0346423
      durun1 |   2.510934   1.653752     1.52   0.129      -.73036    5.752228
       _cons |   13.43295   4.688575     2.87   0.004     4.243509    22.62239
------------------------------------------------------------------------------

. test durtype durfac durfac2 durtre durtra durdev durexp durcost durdec durun1

 ( 1)  durtype = 0
 ( 2)  durfac = 0
 ( 3)  durfac2 = 0
 ( 4)  durtre = 0
 ( 5)  durtra = 0
 ( 6)  durdev = 0
 ( 7)  durexp = 0
 ( 8)  durcost = 0
 ( 9)  durdec = 0
 (10)  durun1 = 0

           chi2( 10) =   18.55
         Prob > chi2 =    0.0463

. 
. 
. 
. ** Add to the model with MultiPKO every combination of two interactions
. 
. gen un4type=untype4*wartype

. gen un4fac=untype4*factnum

. gen un4fac2=untype4*factnum2

. gen un4tre=untype4*treaty

. gen un4tra=untype4*trnsfcap

. gen un4dev=untype4*develop

. gen un4exp=untype4*exp

. gen un4dur=untype4*wardur

. gen un4cost=untype4*logcost

. gen un4dec=untype4*decade

. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4fac , cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -45.045933                Pseudo R2       =     0.4310

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.631263    .622421    -2.62   0.009    -2.851186   -.4113404
     logcost |  -.4410305   .1244166    -3.54   0.000    -.6848825   -.1971784
      wardur |   .0062259   .0055349     1.12   0.261    -.0046223    .0170741
     factnum |  -1.233114   .7064711    -1.75   0.081    -2.617772     .151544
    factnum2 |   .0570087   .0693312     0.82   0.411    -.0788779    .1928953
    trnsfcap |   .0041914   .0016322     2.57   0.010     .0009924    .0073904
     develop |   .0007866   .0004226     1.86   0.063    -.0000417     .001615
         exp |  -5.803813   3.083149    -1.88   0.060    -11.84667    .2390485
      decade |  -.2983112   .1652889    -1.80   0.071    -.6222715     .025649
      treaty |   2.097178   .8088732     2.59   0.010      .511816    3.682541
     untype4 |   18.83447   2.004811     9.39   0.000     14.90512    22.76383
     un4type |   -18.1626          .        .       .            .           .
      un4fac |   .2704169   .3400336     0.80   0.426    -.3960366    .9368705
       _cons |   8.457895   2.095896     4.04   0.000     4.350014    12.56578
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4tre , cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(11)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -45.094546                Pseudo R2       =     0.4304

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   -1.62996   .6209579    -2.62   0.009    -2.847015    -.412905
     logcost |  -.4442191   .1234191    -3.60   0.000    -.6861161   -.2023221
      wardur |   .0062722   .0055161     1.14   0.256    -.0045391    .0170836
     factnum |  -1.274298   .6860597    -1.86   0.063     -2.61895    .0703545
    factnum2 |   .0644586   .0633472     1.02   0.309    -.0596996    .1886168
    trnsfcap |   .0041963   .0016217     2.59   0.010     .0010179    .0073747
     develop |   .0007882   .0004227     1.86   0.062    -.0000404    .0016167
         exp |  -5.741405    3.05485    -1.88   0.060     -11.7288    .2459904
      decade |  -.2975308   .1656973    -1.80   0.073    -.6222916      .02723
      treaty |   2.085209    .806525     2.59   0.010     .5044486    3.665968
     untype4 |   20.14487   1.059523    19.01   0.000     18.06825     22.2215
     un4type |  -18.07088          .        .       .            .           .
       _cons |   8.527563   2.072679     4.11   0.000     4.465188    12.58994
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4tra, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.944226                Pseudo R2       =     0.4323

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.627311   .6219792    -2.62   0.009    -2.846368   -.4082543
     logcost |  -.4367463    .124991    -3.49   0.000    -.6817241   -.1917685
      wardur |   .0060772   .0055728     1.09   0.275    -.0048452    .0169997
     factnum |  -1.158752   .7763602    -1.49   0.136     -2.68039    .3628858
    factnum2 |    .045929   .0837728     0.55   0.584    -.1182627    .2101206
    trnsfcap |   .0041764   .0016595     2.52   0.012      .000924    .0074289
     develop |    .000782   .0004208     1.86   0.063    -.0000427    .0016067
         exp |  -5.874567   3.087768    -1.90   0.057    -11.92648    .1773463
      decade |  -.2965766   .1647771    -1.80   0.072    -.6195338    .0263806
      treaty |   2.103585   .8054965     2.61   0.009     .5248407    3.682329
     untype4 |   20.03621   1.368943    14.64   0.000     17.35313    22.71929
     un4type |   -19.1718          .        .       .            .           .
      un4tra |   .0091593   .0065331     1.40   0.161    -.0036453     .021964
       _cons |   8.314644   2.183752     3.81   0.000     4.034569    12.59472
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4dev, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -45.093397                Pseudo R2       =     0.4304

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.630514   .6215984    -2.62   0.009    -2.848824   -.4122032
     logcost |  -.4439402   .1237085    -3.59   0.000    -.6864044   -.2014761
      wardur |   .0062754   .0055229     1.14   0.256    -.0045493    .0171001
     factnum |   -1.27134   .6836541    -1.86   0.063    -2.611277    .0685976
    factnum2 |   .0637616   .0634378     1.01   0.315    -.0605742    .1880974
    trnsfcap |   .0041972   .0016222     2.59   0.010     .0010178    .0073766
     develop |   .0007883   .0004229     1.86   0.062    -.0000405    .0016171
         exp |  -5.748091   3.067827    -1.87   0.061    -11.76092    .2647388
      decade |   -.297823   .1658572    -1.80   0.073    -.6228971    .0272511
      treaty |   2.087089   .8096438     2.58   0.010     .5002158    3.673961
     untype4 |   19.91872   3.175521     6.27   0.000     13.69481    26.14262
     un4type |  -18.28459          .        .       .            .           .
      un4dev |   .0008869   .0064236     0.14   0.890    -.0117031    .0134768
       _cons |   8.524096   2.071016     4.12   0.000      4.46498    12.58321
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4exp, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =     917.83
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |  -83.34745   7.918957   -10.53   0.000    -98.86832   -67.82658
     un4type |  -250.1052   17.88469   -13.98   0.000    -285.1586   -215.0519
      un4exp |   3036.101      233.8    12.99   0.000     2577.861     3494.34
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4cost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   367.5084          .        .       .            .           .
     un4type |  -42.75305   1.766269   -24.21   0.000    -46.21487   -39.29123
     un4cost |   -22.8972    .093906  -243.83   0.000    -23.08125   -22.71315
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4dur, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   15.54229          .        .       .            .           .
     un4type |  -37.05453          .        .       .            .           .
      un4dur |   .1636757          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4fac2, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.988569                Pseudo R2       =     0.4318

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.629405   .6222739    -2.62   0.009    -2.849039   -.4097701
     logcost |  -.4384897    .124801    -3.51   0.000    -.6830951   -.1938844
      wardur |   .0061447   .0055557     1.11   0.269    -.0047443    .0170337
     factnum |  -1.191282   .7417144    -1.61   0.108    -2.645015    .2624516
    factnum2 |   .0506306   .0769339     0.66   0.510    -.1001571    .2014183
    trnsfcap |   .0041827   .0016466     2.54   0.011     .0009555    .0074099
     develop |   .0007841   .0004217     1.86   0.063    -.0000424    .0016106
         exp |  -5.846678   3.087553    -1.89   0.058    -11.89817     .204815
      decade |  -.2975108    .164951    -1.80   0.071    -.6208088    .0257872
      treaty |   2.101719   .8070534     2.60   0.009     .5199239    3.683515
     untype4 |   19.12643   1.470238    13.01   0.000     16.24482    22.00804
     un4type |  -18.25383          .        .       .            .           .
     un4fac2 |   .0407581   .0351042     1.16   0.246    -.0280449    .1095611
       _cons |   8.378096   2.139534     3.92   0.000     4.184687    12.57151
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4type 
> un4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   101.8966          .        .       .            .           .
     un4type |  -31.46601    2.05997   -15.27   0.000    -35.50348   -27.42854
      un4dec |  -17.01158   .3015785   -56.41   0.000    -17.60267    -16.4205
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac u
> n4tre, cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      41.77
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.616112                Pseudo R2       =     0.4238

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.744727   .6084625    -2.87   0.004    -2.937291   -.5521618
     logcost |  -.4421318   .1265421    -3.49   0.000    -.6901497   -.1941139
      wardur |    .006367   .0056636     1.12   0.261    -.0047334    .0174675
     factnum |  -1.223312   .7261509    -1.68   0.092    -2.646542    .1999176
    factnum2 |    .055091   .0717775     0.77   0.443    -.0855903    .1957723
    trnsfcap |   .0041238   .0016056     2.57   0.010     .0009768    .0072707
     develop |    .000788   .0004273     1.84   0.065    -.0000496    .0016255
         exp |   -6.02479   3.081381    -1.96   0.051    -12.06419    .0146057
      decade |   -.298727   .1685748    -1.77   0.076    -.6291276    .0316735
      treaty |    2.13074   .8230313     2.59   0.010     .5176283    3.743852
     untype4 |   1.911142   1.840373     1.04   0.299    -1.695924    5.518207
      un4fac |   .2419244   .3040255     0.80   0.426    -.3539547    .8378034
       _cons |   8.541413   2.184625     3.91   0.000     4.259627     12.8232
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac u
> n4tra, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      42.94
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.540591                Pseudo R2       =     0.4248

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.718823   .6115488    -2.81   0.005    -2.917436   -.5202093
     logcost |  -.4445983    .127096    -3.50   0.000    -.6937019   -.1954947
      wardur |   .0064933   .0056596     1.15   0.251    -.0045994    .0175859
     factnum |  -1.263972   .7176663    -1.76   0.078    -2.670572    .1426278
    factnum2 |   .0597339   .0692695     0.86   0.389    -.0760318    .1954995
    trnsfcap |   .0042101   .0016325     2.58   0.010     .0010105    .0074098
     develop |   .0007911   .0004288     1.84   0.065    -.0000493    .0016314
         exp |  -5.897876   3.093546    -1.91   0.057    -11.96112    .1653638
      decade |  -.3007372   .1680108    -1.79   0.073    -.6300323    .0285579
      treaty |   2.122607   .8236025     2.58   0.010     .5083758    3.736838
     untype4 |   1.010006   1.857854     0.54   0.587     -2.63132    4.651332
      un4fac |   .5521025   .3990796     1.38   0.167     -.230079    1.334284
      un4tra |  -.0075604   .0070626    -1.07   0.284    -.0214029    .0062821
       _cons |   8.604438    2.17182     3.96   0.000      4.34775    12.86113
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac u
> n4dev, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   98.84353          .        .       .            .           .
      un4fac |   23.13345          .        .       .            .           .
      un4dev |  -.4532804          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac u
> n4exp, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      43.15
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.606758                Pseudo R2       =     0.4240

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.747977   .6082806    -2.87   0.004    -2.940185   -.5557693
     logcost |  -.4429381   .1265533    -3.50   0.000     -.690978   -.1948981
      wardur |   .0063094   .0057116     1.10   0.269    -.0048851     .017504
     factnum |  -1.223153   .7265117    -1.68   0.092     -2.64709    .2007839
    factnum2 |   .0553268   .0715539     0.77   0.439    -.0849164    .1955699
    trnsfcap |   .0041021   .0016182     2.54   0.011     .0009306    .0072737
     develop |   .0007867   .0004277     1.84   0.066    -.0000517     .001625
         exp |  -6.085145    3.11976    -1.95   0.051    -12.19976    .0294724
      decade |  -.2977792   .1690661    -1.76   0.078    -.6291427    .0335843
      treaty |    2.13513   .8228271     2.59   0.009     .5224184    3.747841
     untype4 |   1.238053   2.107407     0.59   0.557    -2.892389    5.368496
      un4fac |   .2627579   .2962626     0.89   0.375    -.3179061    .8434218
      un4exp |   5.248296   12.52295     0.42   0.675    -19.29624    29.79284
       _cons |   8.558239   2.190195     3.91   0.000     4.265535    12.85094
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac  
> un4cost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      47.94
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.514976                Pseudo R2       =     0.4251

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.737728   .6049699    -2.87   0.004    -2.923447   -.5520088
     logcost |  -.4355535   .1274015    -3.42   0.001    -.6852558   -.1858512
      wardur |   .0061961   .0057515     1.08   0.281    -.0050766    .0174689
     factnum |  -1.141833   .7969124    -1.43   0.152    -2.703753    .4200864
    factnum2 |   .0449975   .0841651     0.53   0.593     -.119963     .209958
    trnsfcap |   .0040792    .001642     2.48   0.013     .0008609    .0072976
     develop |    .000783    .000424     1.85   0.065    -.0000481    .0016141
         exp |  -6.004411   3.087435    -1.94   0.052    -12.05567    .0468502
      decade |  -.2956529   .1683917    -1.76   0.079    -.6256946    .0343889
      treaty |   2.120075   .8171818     2.59   0.009     .5184286    3.721722
     untype4 |   17.60729   16.54386     1.06   0.287    -14.81808    50.03267
      un4fac |  -.0704533   .5758297    -0.12   0.903    -1.199059    1.058152
     un4cost |  -.9666936   1.005721    -0.96   0.336    -2.937871    1.004484
       _cons |     8.3329   2.294051     3.63   0.000     3.836643    12.82916
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac u
> n4dur, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   81.27928          .        .       .            .           .
      un4fac |  -24.34519   .4687945   -51.93   0.000    -25.26401   -23.42637
      un4dur |   .7191001   .0254666    28.24   0.000     .6691866    .7690137
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4fac u
> n4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   400.9663          .        .       .            .           .
      un4fac |  -16.07903          .        .       .            .           .
      un4dec |  -67.10745          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre u
> n4tra, cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      41.85
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.637675                Pseudo R2       =     0.4236

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.731872   .6117812    -2.83   0.005    -2.930941    -.532803
     logcost |  -.4466341   .1273262    -3.51   0.000    -.6961888   -.1970794
      wardur |   .0064488   .0056741     1.14   0.256    -.0046722    .0175698
     factnum |  -1.280194   .6982993    -1.83   0.067    -2.648835    .0884477
    factnum2 |   .0647252   .0642776     1.01   0.314    -.0612566     .190707
    trnsfcap |   .0041217   .0016067     2.57   0.010     .0009727    .0072707
     develop |   .0007922   .0004289     1.85   0.065    -.0000485    .0016329
         exp |  -5.961017   3.096915    -1.92   0.054    -12.03086    .1088257
      decade |  -.2996267   .1691915    -1.77   0.077    -.6312358    .0319825
      treaty |   2.119257   .8221416     2.58   0.010     .5078891    3.730625
     untype4 |    3.30621   1.237616     2.67   0.008     .8805259    5.731893
      un4tra |  -.0019387   .0053409    -0.36   0.717    -.0124067    .0085293
       _cons |   8.644863   2.161125     4.00   0.000     4.409136    12.88059
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre u
> n4dev, cluster(cluster)

note: un4tre dropped due to collinearity
Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -50.177735
Iteration 2:   log pseudo-likelihood = -46.156925
Iteration 3:   log pseudo-likelihood =  -45.45674
Iteration 4:   log pseudo-likelihood = -45.379734
Iteration 5:   log pseudo-likelihood = -45.372149
Iteration 6:   log pseudo-likelihood = -45.372023
Iteration 7:   log pseudo-likelihood = -45.372023

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      42.18
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.372023                Pseudo R2       =     0.4269

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.667365   .6179613    -2.70   0.007    -2.878547   -.4561832
     logcost |  -.4487783   .1249424    -3.59   0.000     -.693661   -.2038956
      wardur |   .0064442   .0055766     1.16   0.248    -.0044858    .0173741
     factnum |  -1.305887   .6795787    -1.92   0.055    -2.637837    .0260624
    factnum2 |   .0698309   .0606252     1.15   0.249    -.0489923    .1886541
    trnsfcap |   .0041851   .0016226     2.58   0.010     .0010049    .0073653
     develop |   .0007938   .0004257     1.86   0.062    -.0000407    .0016282
         exp |  -5.766365   3.067678    -1.88   0.060     -11.7789    .2461732
      decade |  -.2991061   .1676298    -1.78   0.074    -.6276544    .0294423
      treaty |   2.090242    .816015     2.56   0.010     .4908816    3.689602
     untype4 |    8.56249   3.750564     2.28   0.022      1.21152    15.91346
      un4dev |  -.0117432   .0076635    -1.53   0.125    -.0267634    .0032769
       _cons |   8.632358   2.100309     4.11   0.000     4.515827    12.74889
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre  
> un4exp, cluster(cluster)

note: un4tre dropped due to collinearity
Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -50.139036
Iteration 2:   log pseudo-likelihood = -46.264137
Iteration 3:   log pseudo-likelihood = -45.663726
Iteration 4:   log pseudo-likelihood = -45.644236
Iteration 5:   log pseudo-likelihood = -45.644205

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      41.79
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.644205                Pseudo R2       =     0.4235

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.744863   .6085802    -2.87   0.004    -2.937658   -.5520675
     logcost |   -.445207   .1255594    -3.55   0.000    -.6912989    -.199115
      wardur |   .0063412   .0056978     1.11   0.266    -.0048264    .0175087
     factnum |  -1.258015   .7029272    -1.79   0.074    -2.635727    .1196972
    factnum2 |   .0618541   .0652136     0.95   0.343    -.0659621    .1896703
    trnsfcap |   .0040724   .0016025     2.54   0.011     .0009316    .0072133
     develop |   .0007891   .0004278     1.84   0.065    -.0000493    .0016276
         exp |  -6.060958    3.10948    -1.95   0.051    -12.15543     .033512
      decade |  -.2980642   .1694643    -1.76   0.079    -.6302081    .0340798
      treaty |   2.126861   .8206104     2.59   0.010     .5184943    3.735228
     untype4 |   2.762737   1.913713     1.44   0.149    -.9880706    6.513545
      un4exp |   3.416569   13.84957     0.25   0.805    -23.72809    30.56122
       _cons |   8.617042   2.158107     3.99   0.000     4.387231    12.84685
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre u
> n4cost, cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      47.36
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood =  -45.51576                Pseudo R2       =     0.4251

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   -1.73823   .6040909    -2.88   0.004    -2.922226   -.5542334
     logcost |  -.4354956   .1274334    -3.42   0.001    -.6852606   -.1857307
      wardur |    .006206   .0057318     1.08   0.279    -.0050282    .0174401
     factnum |  -1.140992   .8024934    -1.42   0.155     -2.71385    .4318666
    factnum2 |   .0446327   .0851964     0.52   0.600    -.1223493    .2116147
    trnsfcap |   .0040875   .0016255     2.51   0.012     .0009016    .0072735
     develop |    .000783   .0004242     1.85   0.065    -.0000483    .0016143
         exp |  -6.001979   3.089497    -1.94   0.052    -12.05728    .0533242
      decade |  -.2957999   .1682979    -1.76   0.079    -.6256578    .0340579
      treaty |   2.120711   .8171255     2.60   0.009     .5191743    3.722247
     untype4 |   16.25452   10.02774     1.62   0.105    -3.399486    35.90852
     un4cost |  -.8987581   .7041306    -1.28   0.202    -2.278829    .4813125
       _cons |   8.332559   2.299968     3.62   0.000     3.824704    12.84041
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre u
> n4dur, cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      53.21
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.902028                Pseudo R2       =     0.4329

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.666386   .6058667    -2.75   0.006    -2.853863   -.4789092
     logcost |  -.4371467    .125279    -3.49   0.000    -.6826891   -.1916044
      wardur |   .0055304   .0058193     0.95   0.342    -.0058752     .016936
     factnum |  -1.045376   .8994968    -1.16   0.245    -2.808357    .7176056
    factnum2 |   .0322828   .1040289     0.31   0.756    -.1716101    .2361756
    trnsfcap |   .0040839   .0017085     2.39   0.017     .0007353    .0074325
     develop |   .0007643   .0004192     1.82   0.068    -.0000574     .001586
         exp |  -6.215339   3.065318    -2.03   0.043    -12.22325   -.2074256
      decade |   -.283677   .1687572    -1.68   0.093    -.6144351    .0470811
      treaty |   2.126205   .8017543     2.65   0.008     .5547954    3.697614
     untype4 |   .2616384   1.392347     0.19   0.851    -2.467311    2.990588
      un4dur |   .0372672   .0110637     3.37   0.001     .0155826    .0589517
       _cons |   8.189294    2.35083     3.48   0.000     3.581751    12.79684
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre u
> n4fac2, cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =      42.05
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.576566                Pseudo R2       =     0.4243

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.743809   .6072485    -2.87   0.004    -2.933994   -.5536242
     logcost |  -.4400876   .1268773    -3.47   0.001    -.6887625   -.1914126
      wardur |   .0063266   .0056811     1.11   0.265    -.0048081    .0174613
     factnum |  -1.189368    .760053    -1.56   0.118    -2.679045    .3003082
    factnum2 |   .0497516   .0789174     0.63   0.528    -.1049236    .2044268
    trnsfcap |   .0041362   .0016218     2.55   0.011     .0009575    .0073149
     develop |   .0007861   .0004266     1.84   0.065    -.0000501    .0016223
         exp |  -6.031379   3.084644    -1.96   0.051    -12.07717    .0144122
      decade |  -.2980732   .1682948    -1.77   0.077    -.6279249    .0317786
      treaty |   2.132655   .8222462     2.59   0.009     .5210824    3.744228
     untype4 |   2.155153   1.374908     1.57   0.117    -.5396165    4.849923
     un4fac2 |   .0366313   .0320368     1.14   0.253    -.0261597    .0994223
       _cons |    8.47297   2.225606     3.81   0.000     4.110862    12.83508
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tre u
> n4dec, cluster(cluster) nolog
note: un4tre dropped due to collinearity

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(11)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.772733                Pseudo R2       =     0.4345

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.642601    .608162    -2.70   0.007    -2.834577   -.4506258
     logcost |  -.4340288   .1247606    -3.48   0.001    -.6785551   -.1895025
      wardur |   .0054638   .0057984     0.94   0.346    -.0059009    .0168286
     factnum |  -1.027766   .9117079    -1.13   0.260     -2.81468    .7591488
    factnum2 |   .0305255   .1066703     0.29   0.775    -.1785445    .2395954
    trnsfcap |   .0040771   .0017175     2.37   0.018     .0007109    .0074433
     develop |   .0007626   .0004169     1.83   0.067    -.0000545    .0015796
         exp |  -6.147222   3.056711    -2.01   0.044    -12.13827   -.1561787
      decade |  -.2829486   .1676093    -1.69   0.091    -.6114567    .0455595
      treaty |   2.112706   .7969894     2.65   0.008     .5506359    3.674777
     untype4 |    97.7463          .        .       .            .           .
      un4dec |  -19.33963   .2403587   -80.46   0.000    -19.81072   -18.86853
       _cons |   8.104858   2.362025     3.43   0.001     3.475375    12.73434
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tra u
> n4dev, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =     936.56
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   169.2352    8.82478    19.18   0.000      151.939    186.5315
      un4tra |   .2651836   .0165985    15.98   0.000     .2326512    .2977161
      un4dev |  -.4134812   .0222142   -18.61   0.000    -.4570202   -.3699422
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tra u
> n4exp, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      42.45
                                                  Prob > chi2     =     0.0001
Log pseudo-likelihood = -45.621886                Pseudo R2       =     0.4238

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.732597   .6115006    -2.83   0.005    -2.931116   -.5340782
     logcost |  -.4481076   .1274012    -3.52   0.000    -.6978093   -.1984059
      wardur |   .0063889   .0056997     1.12   0.262    -.0047824    .0175601
     factnum |  -1.286048   .6932459    -1.86   0.064    -2.644785    .0726893
    factnum2 |   .0661937   .0632809     1.05   0.296    -.0578345     .190222
    trnsfcap |   .0040943   .0016131     2.54   0.011     .0009326     .007256
     develop |   .0007913   .0004291     1.84   0.065    -.0000497    .0016322
         exp |  -6.023827    3.11536    -1.93   0.053    -12.12982    .0821675
      decade |  -.2985776   .1696547    -1.76   0.078    -.6310948    .0339395
      treaty |   2.121751   .8211585     2.58   0.010     .5123099    3.731192
     untype4 |   2.628404   1.907583     1.38   0.168     -1.11039    6.367198
      un4tra |  -.0032371    .005805    -0.56   0.577    -.0146147    .0081405
      un4exp |    7.29548   15.09105     0.48   0.629    -22.28243    36.87339
       _cons |   8.673539    2.15707     4.02   0.000     4.445761    12.90132
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tra u
> n4cost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =     980.44
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   991.9717   51.31419    19.33   0.000     891.3977    1092.546
      un4tra |  -.6373868   .0326793   -19.50   0.000     -.701437   -.5733365
     un4cost |  -63.64041   3.314345   -19.20   0.000     -70.1364   -57.14441
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tra u
> n4dur, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   16.74829          .        .       .            .           .
      un4tra |  -.6438663          .        .       .            .           .
      un4dur |   .8707243          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tra u
> n4fac2, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      43.88
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.387946                Pseudo R2       =     0.4267

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.701607   .6105286    -2.79   0.005    -2.898221   -.5049924
     logcost |  -.4421959   .1266439    -3.49   0.000    -.6904133   -.1939785
      wardur |   .0064039   .0056261     1.14   0.255    -.0046231    .0174308
     factnum |  -1.233738   .7424754    -1.66   0.097    -2.688963    .2214874
    factnum2 |   .0545988   .0754027     0.72   0.469    -.0931878    .2023854
    trnsfcap |   .0042397   .0016404     2.58   0.010     .0010247    .0074548
     develop |   .0007873   .0004271     1.84   0.065    -.0000498    .0016244
         exp |  -5.888204   3.093743    -1.90   0.057    -11.95183    .1754216
      decade |  -.2993992    .167095    -1.79   0.073    -.6268993    .0281009
      treaty |   2.121164   .8206895     2.58   0.010     .5126424    3.729686
     untype4 |   1.522067   1.355403     1.12   0.261    -1.134474    4.178609
      un4tra |  -.0182113   .0103144    -1.77   0.077    -.0384271    .0020045
     un4fac2 |   .1227223   .0609916     2.01   0.044      .003181    .2422636
       _cons |   8.530911   2.188305     3.90   0.000     4.241913    12.81991
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4tra u
> n4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =     967.45
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   322.1743    17.5146    18.39   0.000     287.8463    356.5023
      un4tra |  -.3552783   .0213786   -16.62   0.000    -.3971797    -.313377
      un4dec |   -60.2027   3.303196   -18.23   0.000    -66.67684   -53.72855
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dev u
> n4exp, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   555.8051    26.8941    20.67   0.000     503.0936    608.5165
      un4dev |  -1.827601   .0494067   -36.99   0.000    -1.924437   -1.730766
      un4exp |   3363.676          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dev u
> n4cost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   428.3292          .        .       .            .           .
      un4dev |  -.3089537   .0143806   -21.48   0.000     -.337139   -.2807683
     un4cost |  -19.32878    .565778   -34.16   0.000    -20.43769   -18.21988
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 5 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dev u
> n4dur, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =    1063.44
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   18.25018   1.621326    11.26   0.000     15.07244    21.42792
      un4dev |   -.076403   .0041378   -18.46   0.000    -.0845129    -.068293
      un4dur |   .1884122   .0132777    14.19   0.000     .1623884     .214436
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 3 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dev u
> n4fac2, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =     964.77
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   88.18549   5.903267    14.94   0.000      76.6153    99.75568
      un4dev |  -.2727639    .017876   -15.26   0.000    -.3078002   -.2377276
     un4fac2 |   1.410126   .0977328    14.43   0.000     1.218573    1.601679
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dev u
> n4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   152.9645          .        .       .            .           .
      un4dev |  -.0713038          .        .       .            .           .
      un4dec |  -26.49322          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4exp u
> n4cost, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(12)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   1707.589   13.33983   128.01   0.000     1681.444    1733.735
      un4exp |    2665.03          .        .       .            .           .
     un4cost |  -135.1051   .9188764  -147.03   0.000     -136.906   -133.3041
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 4 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4exp u
> n4dur, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   31.61486          .        .       .            .           .
      un4exp |  -591.8811          .        .       .            .           .
      un4dur |   .4440711          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4exp u
> n4fac2, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      43.57
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.562502                Pseudo R2       =     0.4245

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.746747   .6070391    -2.88   0.004    -2.936522   -.5569725
     logcost |  -.4413738   .1270059    -3.48   0.001    -.6903008   -.1924467
      wardur |   .0062614   .0057232     1.09   0.274    -.0049558    .0174786
     factnum |  -1.191409   .7588321    -1.57   0.116    -2.678693    .2958745
    factnum2 |   .0502687   .0783343     0.64   0.521    -.1032638    .2038011
    trnsfcap |   .0041159   .0016311     2.52   0.012     .0009189    .0073128
     develop |   .0007846   .0004271     1.84   0.066    -.0000525    .0016217
         exp |  -6.100058   3.119679    -1.96   0.051    -12.21452    .0144012
      decade |  -.2969475   .1688712    -1.76   0.079    -.6279289    .0340339
      treaty |    2.13816   .8221832     2.60   0.009     .5267106     3.74961
     untype4 |   1.337352   1.804735     0.74   0.459    -2.199864    4.874567
      un4exp |   6.736891    12.5546     0.54   0.592    -17.86967    31.34345
     un4fac2 |   .0401387    .032113     1.25   0.211    -.0228016     .103079
       _cons |   8.499256   2.231701     3.81   0.000     4.125202    12.87331
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4exp u
> n4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   275.0577          .        .       .            .           .
      un4exp |  -445.4428          .        .       .            .           .
      un4dec |  -48.58906          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4cost 
> un4dur, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   125.5459          .        .       .            .           .
     un4cost |  -10.24061          .        .       .            .           .
      un4dur |   .3196359          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4cost 
> un4fac2, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =      49.74
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -45.513243                Pseudo R2       =     0.4251

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |   -1.73848   .6043346    -2.88   0.004    -2.922954   -.5540063
     logcost |  -.4353525   .1275411    -3.41   0.001    -.6853285   -.1853766
      wardur |   .0062207   .0057489     1.08   0.279     -.005047    .0174884
     factnum |  -1.137923   .8084014    -1.41   0.159     -2.72236    .4465146
    factnum2 |   .0437638   .0867369     0.50   0.614    -.1262375    .2137651
    trnsfcap |   .0041024   .0016467     2.49   0.013     .0008749    .0073299
     develop |    .000783   .0004243     1.85   0.065    -.0000486    .0016146
         exp |  -5.997963   3.089679    -1.94   0.052    -12.05362    .0576967
      decade |  -.2960339   .1682895    -1.76   0.079    -.6258753    .0338075
      treaty |   2.121841   .8183823     2.59   0.010     .5178411    3.725841
     untype4 |   13.84948   13.94786     0.99   0.321    -13.48783    41.18679
     un4cost |  -.7577874   .9085068    -0.83   0.404    -2.538428    1.022853
     un4fac2 |   .0128131   .0543376     0.24   0.814    -.0936866    .1193128
       _cons |   8.328857   2.308168     3.61   0.000     3.804931    12.85278
------------------------------------------------------------------------------

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4cost 
> un4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   295.4186          .        .       .            .           .
     un4cost |  -8.284351          .        .       .            .           .
      un4dec |  -37.46578          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dur u
> n4fac2, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(10)   =          .
                                                  Prob > chi2     =          .
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   37.22077          .        .       .            .           .
      un4dur |   1.064757          .        .       .            .           .
     un4fac2 |   -3.62424          .        .       .            .           .
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 1 failure and 6 successes completely determined.

. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop exp decade treaty untype4 un4dur u
> n4dec, cluster(cluster) nolog

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(13)   =     975.62
                                                  Prob > chi2     =     0.0000
Log pseudo-likelihood = -44.570317                Pseudo R2       =     0.4370

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -1.600592    .612111    -2.61   0.009    -2.800308   -.4008767
     logcost |  -.4272253   .1238022    -3.45   0.001    -.6698732   -.1845774
      wardur |   .0053967   .0057355     0.94   0.347    -.0058448    .0166382
     factnum |   -1.02715   .8661944    -1.19   0.236     -2.72486    .6705594
    factnum2 |   .0324715   .0989898     0.33   0.743     -.161545     .226488
    trnsfcap |   .0040333   .0016983     2.37   0.018     .0007046    .0073619
     develop |   .0007615   .0004126     1.85   0.065    -.0000471    .0015701
         exp |  -5.974324   3.058439    -1.95   0.051    -11.96875    .0201054
      decade |  -.2829107   .1653055    -1.71   0.087    -.6069036    .0410821
      treaty |   2.079662   .7888262     2.64   0.008     .5335912    3.625733
     untype4 |   967.4221   51.61562    18.74   0.000     866.2574    1068.587
      un4dur |  -1.286904   .0740908   -17.37   0.000     -1.43212   -1.141689
      un4dec |  -187.0283   9.974618   -18.75   0.000    -206.5782   -167.4784
       _cons |   7.978325   2.297419     3.47   0.001     3.475466    12.48118
------------------------------------------------------------------------------

note: 0 failures and 4 successes completely determined.

. 
. 
. 
. ** Equality of Means Tests
. ** Comparisons of means of covariates in the "treatment group" (UN intervention) to the "control group" 
> (no UN mission)
. 
. ** Start with MultiPKO as the treatment
. 
. sort untype4

. by untype4: sum wartype logcost wardur factnum factnum2 trnsfcap  treaty develop exp decade

_______________________________________________________________________________
-> untype4 = 0

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
     wartype |       115    .6521739    .4783649          0          1
     logcost |       115    11.80511    2.429621   6.907755   15.67181
      wardur |       115        75.4    91.88177          1        600
     factnum |       115    3.252174    1.450006          2         11
    factnum2 |       115    12.66087    13.97422          4        121
-------------+--------------------------------------------------------
    trnsfcap |       115     57.5817     174.655       -112   1272.469
      treaty |       115    .2434783    .4310596          0          1
     develop |       115    556.7972    878.9112         10       5387
         exp |       115    .1241705    .0849891       .013       .508
      decade |       115    3.443478    1.458137          1          6

_______________________________________________________________________________
-> untype4 = 1

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
     wartype |         7    .4285714    .5345225          0          1
     logcost |         7    13.52207    1.662663   10.91509   15.20181
      wardur |         7    148.1429    99.34188         12        288
     factnum |         7    4.714286     2.21467          2          8
    factnum2 |         7    26.42857    22.49339          4         64
-------------+--------------------------------------------------------
    trnsfcap |         7    76.50767    122.2171   8.300904   344.5972
      treaty |         7           1           0          1          1
     develop |         7    312.5039    214.7017         18        584
         exp |         7    .1044591    .0472903       .033   .1577578
      decade |         7           4    .8164966          3          5


. 
. ttest wartype , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    .6521739    .0446078    .4783649    .5638063    .7405415
       1 |       7    .4285714    .2020305    .5345225   -.0657794    .9229223
---------+--------------------------------------------------------------------
combined |     122    .6393443    .0436537    .4821709    .5529202    .7257683
---------+--------------------------------------------------------------------
    diff |            .2236025    .1873801               -.1473971     .594602
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   1.1933                t =   1.1933              t =   1.1933
   P < t =   0.8824          P > |t| =   0.2351          P > t =   0.1176

. ttest logcost , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    11.80511    .2265634    2.429621    11.35629    12.25393
       1 |       7    13.52207    .6284277    1.662663    11.98436    15.05978
---------+--------------------------------------------------------------------
combined |     122    11.90362    .2191526     2.42062    11.46975    12.33749
---------+--------------------------------------------------------------------
    diff |           -1.716962    .9331892               -3.564612    .1306876
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -1.8399                t =  -1.8399              t =  -1.8399
   P < t =   0.0341          P > |t| =   0.0683          P > t =   0.9659

. ttest wardur , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115        75.4    8.568019    91.88177    58.42682    92.37318
       1 |       7    148.1429     37.5477    99.34188    56.26694    240.0188
---------+--------------------------------------------------------------------
combined |     122    79.57377    8.460024    93.44401    62.82492    96.32262
---------+--------------------------------------------------------------------
    diff |           -72.74286    35.92016               -143.8623    -1.62344
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -2.0251                t =  -2.0251              t =  -2.0251
   P < t =   0.0225          P > |t| =   0.0451          P > t =   0.9775

. ttest factnum , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    3.252174    .1352137    1.450006    2.984317    3.520031
       1 |       7    4.714286    .8370665     2.21467    2.666058    6.762514
---------+--------------------------------------------------------------------
combined |     122    3.336066    .1385129    1.529925    3.061843    3.610288
---------+--------------------------------------------------------------------
    diff |           -1.462112    .5829895                -2.61639   -.3078332
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -2.5080                t =  -2.5080              t =  -2.5080
   P < t =   0.0067          P > |t| =   0.0135          P > t =   0.9933

. ttest treaty , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    .2434783    .0401965    .4310596    .1638493    .3231072
       1 |       7           1           0           0           1           1
---------+--------------------------------------------------------------------
combined |     122    .2868852    .0411189    .4541727    .2054796    .3682909
---------+--------------------------------------------------------------------
    diff |           -.7565217    .1635615               -1.080362   -.4326814
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -4.6253                t =  -4.6253              t =  -4.6253
   P < t =   0.0000          P > |t| =   0.0000          P > t =   1.0000

. ttest develop, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    556.7972    81.95889    878.9112    394.4373    719.1571
       1 |       7    312.5039     81.1496    214.7017     113.938    511.0698
---------+--------------------------------------------------------------------
combined |     122    542.7804    77.53031    856.3503    389.2887    696.2721
---------+--------------------------------------------------------------------
    diff |            244.2933    334.0179               -417.0388    905.6254
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.7314                t =   0.7314              t =   0.7314
   P < t =   0.7670          P > |t| =   0.4660          P > t =   0.2330

. ttest exp, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    .1241705    .0079253    .0849891    .1084705    .1398704
       1 |       7    .1044591    .0178741    .0472903    .0607228    .1481953
---------+--------------------------------------------------------------------
combined |     122    .1230395    .0075408    .0832908    .1081105    .1379685
---------+--------------------------------------------------------------------
    diff |            .0197114      .03251               -.0446561    .0840789
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.6063                t =   0.6063              t =   0.6063
   P < t =   0.7273          P > |t| =   0.5454          P > t =   0.2727

. ttest trnsfcap , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115     57.5817    16.28666     174.655    25.31796    89.84545
       1 |       7    76.50767    46.19374    122.2171   -36.52433    189.5397
---------+--------------------------------------------------------------------
combined |     122    58.66762    15.54999    171.7552    27.88232    89.45292
---------+--------------------------------------------------------------------
    diff |           -18.92596    67.11972               -151.8183    113.9664
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.2820                t =  -0.2820              t =  -0.2820
   P < t =   0.3892          P > |t| =   0.7784          P > t =   0.6108

. ttest decade , by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |     115    3.443478     .135972    1.458137    3.174119    3.712838
       1 |       7           4    .3086067    .8164966    3.244867    4.755133
---------+--------------------------------------------------------------------
combined |     122     3.47541    .1297257    1.432868    3.218584    3.732236
---------+--------------------------------------------------------------------
    diff |           -.5565217    .5578229               -1.660972    .5479289
------------------------------------------------------------------------------
Degrees of freedom: 120

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.9977                t =  -0.9977              t =  -0.9977
   P < t =   0.1602          P > |t| =   0.3204          P > t =   0.8398

. 
. * Results show that about half of our covariates have statistically different mean values in the two gro
> ups
. 
. * Re-run tests, restricting the comparisons to cases where treaty==1 
. * Then do the same for unintrvn instead of untype4
. 
. by untype4: sum wartype logcost wardur factnum factnum2 trnsfcap  develop exp decade if treaty==1

_______________________________________________________________________________
-> untype4 = 0

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
     wartype |        28    .6428571      .48795          0          1
     logcost |        28    12.74758    2.001337   8.006368   15.07568
      wardur |        28    113.2143    131.0975          4        600
     factnum |        28    3.642857    2.076576          2         11
    factnum2 |        28    17.42857    23.51269          4        121
-------------+--------------------------------------------------------
    trnsfcap |        28    142.5732    316.2972       -112   1272.469
     develop |        28    958.1937    1432.988         10       5387
         exp |        28    .1487011    .1160677       .018       .508
      decade |        28    3.928571    1.488936          1          6

_______________________________________________________________________________
-> untype4 = 1

    Variable |       Obs        Mean    Std. Dev.       Min        Max
-------------+--------------------------------------------------------
     wartype |         7    .4285714    .5345225          0          1
     logcost |         7    13.52207    1.662663   10.91509   15.20181
      wardur |         7    148.1429    99.34188         12        288
     factnum |         7    4.714286     2.21467          2          8
    factnum2 |         7    26.42857    22.49339          4         64
-------------+--------------------------------------------------------
    trnsfcap |         7    76.50767    122.2171   8.300904   344.5972
     develop |         7    312.5039    214.7017         18        584
         exp |         7    .1044591    .0472903       .033   .1577578
      decade |         7           4    .8164966          3          5


. 
. 
. ttest wartype if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    .6428571    .0922139      .48795    .4536499    .8320644
       1 |       7    .4285714    .2020305    .5345225   -.0657794    .9229223
---------+--------------------------------------------------------------------
combined |      35          .6    .0840168    .4970501    .4292573    .7707427
---------+--------------------------------------------------------------------
    diff |            .2142857    .2099121               -.2127836     .641355
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   1.0208                t =   1.0208              t =   1.0208
   P < t =   0.8426          P > |t| =   0.3148          P > t =   0.1574

. ttest logcost if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    12.74758    .3782172    2.001337    11.97155    13.52362
       1 |       7    13.52207    .6284277    1.662663    11.98436    15.05978
---------+--------------------------------------------------------------------
combined |      35    12.90248    .3280836    1.940969    12.23573    13.56923
---------+--------------------------------------------------------------------
    diff |           -.7744855    .8215548               -2.445951    .8969804
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.9427                t =  -0.9427              t =  -0.9427
   P < t =   0.1763          P > |t| =   0.3527          P > t =   0.8237

. ttest wardur if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    113.2143    24.77509    131.0975       62.38    164.0486
       1 |       7    148.1429     37.5477    99.34188    56.26694    240.0188
---------+--------------------------------------------------------------------
combined |      35       120.2    21.10561    124.8625    77.30824    163.0918
---------+--------------------------------------------------------------------
    diff |           -34.92857    53.21125               -143.1877    73.33053
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.6564                t =  -0.6564              t =  -0.6564
   P < t =   0.2581          P > |t| =   0.5161          P > t =   0.7419

. ttest factnum if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    3.642857     .392436    2.076576    2.837645    4.448069
       1 |       7    4.714286    .8370665     2.21467    2.666058    6.762514
---------+--------------------------------------------------------------------
combined |      35    3.857143    .3577306    2.116363    3.130147    4.584139
---------+--------------------------------------------------------------------
    diff |           -1.071429    .8884088                -2.87891    .7360528
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -1.2060                t =  -1.2060              t =  -1.2060
   P < t =   0.1182          P > |t| =   0.2364          P > t =   0.8818

. ttest factnum2 if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    17.42857    4.443481    23.51269    8.311302    26.54584
       1 |       7    26.42857    8.501701    22.49339     5.62566    47.23148
---------+--------------------------------------------------------------------
combined |      35    19.22857    3.933926    23.27342    11.23387    27.22327
---------+--------------------------------------------------------------------
    diff |                  -9    9.859009                -29.0583     11.0583
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.9129                t =  -0.9129              t =  -0.9129
   P < t =   0.1840          P > |t| =   0.3679          P > t =   0.8160

. ttest trnsfcap if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    142.5732    59.77455    316.2972    19.92593    265.2204
       1 |       7    76.50767    46.19374    122.2171   -36.52433    189.5397
---------+--------------------------------------------------------------------
combined |      35    129.3601    48.63904    287.7524    30.51365    228.2065
---------+--------------------------------------------------------------------
    diff |             66.0655    122.8893               -183.9546    316.0856
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.5376                t =   0.5376              t =   0.5376
   P < t =   0.7028          P > |t| =   0.5945          P > t =   0.2972

. ttest develop if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    958.1937    270.8093    1432.988    402.5389    1513.849
       1 |       7    312.5039     81.1496    214.7017     113.938    511.0698
---------+--------------------------------------------------------------------
combined |      35    829.0558    220.8741    1306.709    380.1855    1277.926
---------+--------------------------------------------------------------------
    diff |            645.6898    549.1033               -471.4693    1762.849
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   1.1759                t =   1.1759              t =   1.1759
   P < t =   0.8760          P > |t| =   0.2480          P > t =   0.1240

. ttest exp if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    .1487011    .0219347    .1160677    .1036947    .1937075
       1 |       7    .1044591    .0178741    .0472903    .0607228    .1481953
---------+--------------------------------------------------------------------
combined |      35    .1398527    .0180596    .1068418    .1031512    .1765541
---------+--------------------------------------------------------------------
    diff |             .044242    .0451761               -.0476694    .1361535
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.9793                t =   0.9793              t =   0.9793
   P < t =   0.8327          P > |t| =   0.3345          P > t =   0.1673

. ttest decade if treaty==1, by(untype4)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
       0 |      28    3.928571    .2813825    1.488936    3.351222    4.505921
       1 |       7           4    .3086067    .8164966    3.244867    4.755133
---------+--------------------------------------------------------------------
combined |      35    3.942857    .2317013    1.370763    3.471984    4.413731
---------+--------------------------------------------------------------------
    diff |           -.0714286    .5878327               -1.267383    1.124526
------------------------------------------------------------------------------
Degrees of freedom: 33

                      Ho: mean(0) - mean(1) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.1215                t =  -0.1215              t =  -0.1215
   P < t =   0.4520          P > |t| =   0.9040          P > t =   0.5480

. 
. * No statistically significant differences if we restrict comparison to cases where there was a treaty. 
. * There have been treaties in all multidimensional peacekeeping operations.  
. * So, DS2000 consider the interactive effects of treaty and UN peacekeeping (see conditional effects plo
> ts in DS2000). 
. 
. 
. * Now replace untype4 with unintrvn and re-do this exercise
. 
. sort unintrvn

. 
. ttest wartype , by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    .6395349     .052078    .4829515    .5359899    .7430799
     Yes |      36    .6388889    .0811894    .4871361    .4740657     .803712
---------+--------------------------------------------------------------------
combined |     122    .6393443    .0436537    .4821709    .5529202    .7257683
---------+--------------------------------------------------------------------
    diff |             .000646    .0961131               -.1896512    .1909432
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.0067                t =   0.0067              t =   0.0067
   P < t =   0.5027          P > |t| =   0.9946          P > t =   0.4973

. ttest logcost , by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    11.48504     .273639    2.537624    10.94097    12.02911
     Yes |      36    12.90356     .296224    1.777344    12.30219    13.50493
---------+--------------------------------------------------------------------
combined |     122    11.90362    .2191526     2.42062    11.46975    12.33749
---------+--------------------------------------------------------------------
    diff |           -1.418519    .4648112               -2.338813   -.4982256
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -3.0518                t =  -3.0518              t =  -3.0518
   P < t =   0.0014          P > |t| =   0.0028          P > t =   0.9986

. ttest wardur, by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    80.19767    10.61968    98.48282    59.08291    101.3124
     Yes |      36    78.08333    13.57018    81.42108     50.5344    105.6323
---------+--------------------------------------------------------------------
combined |     122    79.57377    8.460024    93.44401    62.82492    96.32262
---------+--------------------------------------------------------------------
    diff |            2.114341    18.62557                 -34.763    38.99168
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.1135                t =   0.1135              t =   0.1135
   P < t =   0.5451          P > |t| =   0.9098          P > t =   0.4549

. ttest factnum, by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    3.151163    .1625022    1.506983    2.828065    3.474261
     Yes |      36    3.777778     .252326    1.513956    3.265529    4.290027
---------+--------------------------------------------------------------------
combined |     122    3.336066    .1385129    1.529925    3.061843    3.610288
---------+--------------------------------------------------------------------
    diff |            -.626615    .2995536                -1.21971   -.0335197
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -2.0918                t =  -2.0918              t =  -2.0918
   P < t =   0.0193          P > |t| =   0.0386          P > t =   0.9807

. ttest develop, by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    492.7345    94.92656    880.3127    303.9951    681.4739
     Yes |      36    662.3344     132.536    795.2161     393.272    931.3968
---------+--------------------------------------------------------------------
combined |     122    542.7804    77.53031    856.3503    389.2887    696.2721
---------+--------------------------------------------------------------------
    diff |           -169.5999    169.9962               -506.1806    166.9808
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -0.9977                t =  -0.9977              t =  -0.9977
   P < t =   0.1602          P > |t| =   0.3204          P > t =   0.8398

. ttest exp, by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    .1259202    .0088139    .0817368    .1083958    .1434446
     Yes |      36    .1161578    .0146148    .0876885    .0864883    .1458273
---------+--------------------------------------------------------------------
combined |     122    .1230395    .0075408    .0832908    .1081105    .1379685
---------+--------------------------------------------------------------------
    diff |            .0097624    .0165788               -.0230624    .0425872
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =   0.5888                t =   0.5888              t =   0.5888
   P < t =   0.7215          P > |t| =   0.5571          P > t =   0.2785

. ttest decade, by(unintrvn)

Two-sample t test with equal variances

------------------------------------------------------------------------------
   Group |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
      No |      86    3.348837    .1558415    1.445215    3.038983    3.658692
     Yes |      36    3.777778    .2292538    1.375523    3.312368    4.243188
---------+--------------------------------------------------------------------
combined |     122     3.47541    .1297257    1.432868    3.218584    3.732236
---------+--------------------------------------------------------------------
    diff |           -.4289406    .2829225               -.9891074    .1312262
------------------------------------------------------------------------------
Degrees of freedom: 120

                     Ho: mean(No) - mean(Yes) = diff = 0

     Ha: diff < 0               Ha: diff != 0              Ha: diff > 0
       t =  -1.5161                t =  -1.5161              t =  -1.5161
   P < t =   0.0661          P > |t| =   0.1321          P > t =   0.9339

. 
. * Only LOGCOST & FACTNUM have significant differences across the two groups. DECADE also shows weakly si
> gnificant means differences.
. 
. 
. 
. 
. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 */
. /* Figure plots estimates of the marginal effects of MultiPKOs across wars of different duration from th
> e original and modified models  */
. /* The supplement online includes the same sort of figure for other interaction terms */
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. * First, replicate their figure
. 
. * original model
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade,
>  cluster(cluster)

. 
. generate ds0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48

. 
. generate ds1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48

. 
. gen mds=exp(ds1)/(1+exp(ds1))-exp(ds0)/(1+exp(ds0)) 

. 
. label var mds "original model"

. 
. * modified model (with interaction term "un4wdur")
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade 
> un4wdur, cluster(cluster)

. 
. generate kz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48 +_b[un4wdur]*0

. 
. generate kz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48 +_b[un4wdur]*wardur

. 
. gen mkz=exp(kz1)/(1+exp(kz1))-exp(kz0)/(1+exp(kz0))

. 
. label var mkz "modified model"

. 
. graph7 mds mkz wardur if wardur<320, xlabel ylabel connect(ss) saving(C:\Temp\KZfig8a ,replace)

. 
. 
. ** Now add confidence intervals (this must be done in stata version 8)
. ** This is Figure 1 in the ISQ piece
. 
. * original model 
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade,
>  cluster(cluster)

. 
. predictnl mfx=-(exp(_b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3
> .34 + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b
> [exp]*.12 + _b[decade]*3.48)/(1+exp(_b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur
>  + _b[factnum]*3.34 + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[deve
> lop]*542.78 + _b[exp]*.12 + _b[decade]*3.48)) - exp(_b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _
> b[wardur]*wardur + _b[factnum]*3.34 + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[treat
> y]*.29 + _b[develop]*542.78 + _b[exp]*.12 + _b[decade]*3.48)/(1+exp(_b[_cons] + _b[wartype]*.64 + _b[log
> cost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untyp
> e4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]*.12 + _b[decade]*3.48))),ci(lc uc)
note: Confidence intervals calculated using Z critical values

. 
. gr7 mfx lc uc wardur,xlabel ylabel connect(ss) saving(C:\Temp\fig1parta,replace)

. 
. 
. * modified model (with interaction term "un4wdur")
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade 
> un4wdur, cluster(cluster)

. 
. predictnl mfx2=exp(_b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.
> 34 + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[
> exp]*.12 + _b[decade]*3.48 +_b[un4wdur]*wardur)/(1+exp(_b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  
> + _b[wardur]*wardur + _b[factnum]*3.34 + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[tr
> eaty]*.29 + _b[develop]*542.78 + _b[exp]*.12 + _b[decade]*3.48 +_b[un4wdur]*wardur))-exp(_b[_cons] + _b[
> wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 + _b[factnum2]*13.45 + _b[trnsf
> cap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]*.12 + _b[decade]*3.48 +_b[un4
> wdur]*0)/(1+exp(_b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48 +_b[un4wdur]*0)),ci(lc2 uc2)
note: Confidence intervals calculated using Z critical values

. 
. gr7 mfx2 lc2 uc2 wardur,xlabel ylabel connect(ss) saving(C:\Temp\fig1partb,replace)

. 
. * see excel file "Fig 1 with conf intervals" for figures that plot the marginal effects and confidence i
> ntervals
. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 -- replace MultiPKO with Unintrvn */
. /* This is Figure 2 in the supplement */
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty unintrvn exp decade
> , cluster(cluster)

. 
. generate rs0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48

. 
. generate rs1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48

. 
. gen rds=exp(rs1)/(1+exp(rs1))-exp(rs0)/(1+exp(rs0))

. 
. label var rds "original model"

. 
. 
. /*"modified" model (with interaction term "unwdur")*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty unintrvn exp decade
>  unwdur, cluster(cluster)

. 
. generate tz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48 +_b[unwdur]*0

. 
. generate tz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48 +_b[unwdur]*wardur

. 
. gen tkz=exp(tz1)/(1+exp(tz1))-exp(tz0)/(1+exp(tz0))

. 
. label var tkz "modified model"

. 
. 
. graph7 rds tkz wardur if wardur<320, xlabel ylabel connect(ss) saving(C:\Temp\Figure2 ,replace)

. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 -- replace MultiPKO with un2cint */
. /* This is Figure 3 in the supplement */
. 
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. gen un2wdur=un2cint*wardur

. 
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty un2cint exp decade,
>  cluster(cluster)

. 
. generate rs0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[un2cint]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48

. 
. generate rs1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[un2cint]*3 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48

. 
. gen mds=exp(rs1)/(1+exp(rs1))-exp(rs0)/(1+exp(rs0))

. 
. label var mds "original model"

. 
. 
. /*"modified" model (with interaction term "un2wdur")*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty un2cint exp decade 
> un2wdur, cluster(cluster)

. 
. generate tz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[un2cint]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48 +_b[un2wdur]*0

. 
. generate tz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*wardur + _b[factnum]*3.34 +
>  _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[un2cint]*3 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp]
> *.12 + _b[decade]*3.48 +_b[un2wdur]*wardur

. 
. gen mkz=exp(tz1)/(1+exp(tz1))-exp(tz0)/(1+exp(tz0))

. 
. label var mkz "modified model"

. 
. 
. graph7 mds mkz wardur if wardur<320, xlabel ylabel connect(ss) saving(C:\Temp\Figure3 ,replace)

. 
. 
. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 -- MultiPKO interacted with log of deaths and displacem
> ents */
. /* This is Figure 4 in the supplement */
. 
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. gen un4cost=untype4*logcost

. 
. 
. /*" original" model*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade,
>  cluster(cluster)

. 
. generate ds0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48

. 
. generate ds1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48

. 
. gen mds=exp(ds1)/(1+exp(ds1))-exp(ds0)/(1+exp(ds0))

. 
. label var mds "original model"

. 
. 
. /*"modified" model (with interaction term "un4cost")*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade 
> un4cost, cluster(cluster)

. 
. generate kz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48 +_b[un4cost]*0

. 
. generate kz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[exp
> ]*.12 + _b[decade]*3.48 +_b[un4cost]*logcost

. 
. gen mkz=exp(kz1)/(1+exp(kz1))-exp(kz0)/(1+exp(kz0))

. 
. label var mkz "modified model"

. 
. 
. graph7 mds mkz logcost, xlabel ylabel connect(ss) saving(C:\Temp\Figure4 ,replace)

. 
. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 -- Any UN intervention interacted with log of deaths an
> d displacements */
. /* This is Figure 5 in the supplement */
. 
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. gen uncost=unintrvn*logcost

. 
. 
. /*" original" model*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty unintrvn exp decade
> , cluster(cluster)

. 
. generate ds0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48

. 
. generate ds1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48

. 
. gen mds=exp(ds1)/(1+exp(ds1))-exp(ds0)/(1+exp(ds0))

. 
. label var mds "original model"

. 
. 
. /*"modified" model (with interaction term "uncost")*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty unintrvn exp decade
>  uncost, cluster(cluster)

. 
. generate kz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48 +_b[uncost]*0

. 
. generate kz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*logcost  + _b[wardur]*79.57 + _b[factnum]*3.34 
> + _b[factnum2]*13.45 + _b[trnsfcap]*58.67 + _b[unintrvn]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48 +_b[uncost]*logcost

. 
. gen mkz=exp(kz1)/(1+exp(kz1))-exp(kz0)/(1+exp(kz0))

. 
. label var mkz "modified model"

. 
. 
. graph7 mds mkz logcost, xlabel ylabel connect(ss) saving(C:\Temp\Figure5 ,replace)

. 
. 
. 
. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 -- MultiPKO interacted with net current transfers */
. /* This is Figure 6 in the supplement */
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. gen un4tran = untype4 * trnsfcap

. 
. 
. /*" original" model*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade,
>  cluster(cluster)

. 
. generate ds0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48

. 
. generate ds1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48

. 
. gen mds=exp(ds1)/(1+exp(ds1))-exp(ds0)/(1+exp(ds0)) 

. 
. label var mds "original model"

. 
. 
. /*"modified" model (with interaction term "un4tran")*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty untype4 exp decade 
> un4tran , cluster(cluster)

. 
. generate kz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[untype4]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48 +_b[un4tran]*0

. 
. generate kz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[untype4]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[ex
> p]*.12 + _b[decade]*3.48 +_b[un4tran]*trnsfcap

. 
. gen mkz=exp(kz1)/(1+exp(kz1))-exp(kz0)/(1+exp(kz0))

. 
. label var mkz "modified model"

. 
. 
. graph7 mds mkz trnsfcap, xlabel ylabel connect(ss) saving(C:\Temp\Figure6 ,replace)

. 
. 
. 
. 
. /* Do-file to Replicate King and Zeng's Figure 8 -- Any UN intervention interacted with net current tran
> sfers */
. /* This is Figure 7 in the supplement */
. 
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. gen untrans = unintrvn * trnsfcap

. 
. 
. /*" original" model*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty unintrvn exp decade
> , cluster(cluster)

. 
. generate ds0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[e
> xp]*.12 + _b[decade]*3.48

. 
. generate ds1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[e
> xp]*.12 + _b[decade]*3.48

. 
. gen mds=exp(ds1)/(1+exp(ds1))-exp(ds0)/(1+exp(ds0)) 

. 
. label var mds "original model"

. 
. 
. 
. /*"modified" model (with interaction term "untrans")*/
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap develop treaty unintrvn exp decade
>  untrans, cluster(cluster)

. 
. generate kz0 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*0 + _b[treaty]*.29 + _b[develop]*542.78 + _b[e
> xp]*.12 + _b[decade]*3.48 +_b[untrans]*0

. 
. generate kz1 = _b[_cons] + _b[wartype]*.64 + _b[logcost]*11.90  + _b[wardur]*79.57 + _b[factnum]*3.34 + 
> _b[factnum2]*13.45 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*1 + _b[treaty]*.29 + _b[develop]*542.78 + _b[e
> xp]*.12 + _b[decade]*3.48 +_b[untrans]*trnsfcap

. 
. gen mkz=exp(kz1)/(1+exp(kz1))-exp(kz0)/(1+exp(kz0))

. 
. label var mkz "modified model"

. 
. 
. graph7 mds mkz trnsfcap, xlabel ylabel connect(ss) saving(C:\Temp\Figure7 ,replace)

. 
. 
. 
. 
. 
. 
. 
. 
. 
. /* Replicate King and Zeng's Figure 4 */
. /* Figure is supposed to show the sensitivity of counterfactual predictions to small specification chang
> es to the model */
. 
. use "C:\Temp\DS2000replication.dta", clear

. 
. * Estimate the original model
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap untype4 treaty develop exp decade,
>  cluster(cluster) nolog

. 
. 
. * Next, generate predictions from that model
. 
. predict p_orig
(option p assumed; Pr(pbs2s3))

. 
. predict p
(option p assumed; Pr(pbs2s3))

. 
. generate ds = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]*f
> actnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[untype4]*(untype4) + _b[treaty]*treaty + _b[
> develop]*develop + _b[exp]*exp + _b[decade]*decade 

. 
. generate pds=1/(1+exp(-ds))

. 
. corr p_orig pds
(obs=122)

             |   p_orig      pds
-------------+------------------
      p_orig |   1.0000
         pds |   1.0000   1.0000


. 
. 
. * Now estimate the modified model (including interaction between war duration and untype4)
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap untype4 treaty develop exp decade 
> un4wdur, cluster(cluster) nolog

. 
. 
. * Compute probabilities based on that model
. 
. predict p_mod
(option p assumed; Pr(pbs2s3))

. 
. 
. * Plot the two sets of predictions and this gives us the first panel of KZ's figure 4 (In sample fit)
. 
. graph7 p_orig p_mod , saving(C:\Temp\KZfig4a ,replace)

. 
. 
. 
. ** Now plug in the counterfactuals to get the second panel of KZ's figure 4 (counterfactual predictions)
. ** Use the estimated coefficients from the logit models but replace untype4 with 1-untype4 (or un4_count
> ) and replace un4wdur with un4wdur__count
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap untype4 treaty develop exp decade,
>  cluster(cluster) nolog

. 
. generate DS0 = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]*
> factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[untype4]*(1-untype4) + _b[treaty]*treaty + 
> _b[develop]*develop + _b[exp]*exp + _b[decade]*decade 

. 
. generate Pds0=1/(1+exp(-DS0))

. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap untype4 treaty develop exp decade 
> un4wdur, cluster(cluster) nolog

. 
. generate KZ0 = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]*
> factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[untype4]*(1-untype4) + _b[treaty]*treaty + 
> _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[un4wdur]*(un4wdur__count)

. 
. generate Pkz0=1/(1+exp(-KZ0))

. 
. graph7 Pds0 Pkz0 , saving(C:\Temp\KZfig4b ,replace)

. 
. * This is Figure 4 in King and Zeng's paper
. 
. 
. 
. * In our ISQ paper we produce similar graphs for the model with any UN intervention in our Figure 2
. * Figure 2 can be replicated as follows:
. 
. 
. * create interactions with other variables and unintrvn
. 
. gen uncost=unintrvn*logcost

. gen unexp=unintrvn*exp

. gen undev=unintrvn*develop

. gen unfac=unintrvn*factnum

. gen unwar=unintrvn*wartype

. gen undec=unintrvn*decade

. gen unfac2=unintrvn*factnum2

. gen untre=unintrvn*treaty

. gen untra=unintrvn*trnsfcap

. 
. * create interactions with the counterfactual (1-unintrvn) and the other covariates
. 
. gen uncost_c=un_count*logcost

. gen unexp_c=un_count*exp

. gen undev_c=un_count*develop

. gen unfac_c=un_count*factnum

. gen unwar_c=un_count*wartype

. gen undec_c=un_count*decade

. gen unfac2_c=un_count*factnum2

. gen untre_c=un_count*treaty

. gen untra_c=un_count*trnsfcap

. 
. 
. * Estimate "Original" Model with no interactions (original model now controls for unintrvn instead of mu
> ltipko)
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
> , cluster(cluster)

. 
. * Compute probabilities of the counterfactual (use estimates from logit; replace unintrvn with (1-unintr
> vn))
. 
. generate ds1 = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]*
> factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty +
>  _b[develop]*develop + _b[exp]*exp + _b[decade]*decade 

. 
. generate pds1=1/(1+exp(-ds1))

. 
. label var pds1 "Original Model Counterfactual Prediction"

. 
. * Estimate "Modified" Model with UN interaction with war duration
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>   unwdur, cluster(cluster)

. 
. generate KZ3b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[unwdur]* un_count_dur 

. 
. generate Pkz3b=1/(1+exp(-KZ3b))

. 
. label var Pkz3b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz3b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Wardur) saving(C:\Temp\Fig2a , replace)

. 
. 
. * Interaction with LOGCOST
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  uncost , cluster(cluster)

. 
. generate KZ4b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[uncost]*uncost_c 

. 
. generate Pkz4b=1/(1+exp(-KZ4b))

. 
. label var Pkz4b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz4b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Logcost) saving(C:\Temp\Fig2b , replace)

. 
. 
. * Interaction with EXP
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  unexp, cluster(cluster)

. 
. generate KZ5b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[unexp]*unexp_c 

. 
. generate Pkz5b=1/(1+exp(-KZ5b))

. 
. label var Pkz5b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz5b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Exp) saving(C:\Temp\Fig2c , replace) 

. 
. 
. * Interaction with DEVELOP
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  undev, cluster(cluster)

. 
. generate KZ6b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[undev]*undev_c 

. 
. generate Pkz6b=1/(1+exp(-KZ6b))

. 
. label var Pkz6b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz6b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Develop) saving(C:\Temp\Fig2d , replace)

. 
. 
. * Interaction with FACTNUM
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  unfac, cluster(cluster)

. 
. generate KZ7b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[unfac]*unfac_c 

. 
. generate Pkz7b=1/(1+exp(-KZ7b))

. 
. label var Pkz7b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz7b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Factnum) saving(C:\Temp\Fig2e , replace)

. 
. 
. * Interaction with WARTYPE
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  unwar, cluster(cluster)

. 
. generate KZ8b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[unwar]*unwar_c 

. 
. generate Pkz8b=1/(1+exp(-KZ8b))

. 
. label var Pkz8b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz8b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Wartype) saving(C:\Temp\Fig2f , replace)

. 
. 
. * Interaction with DECADE
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  undec, cluster(cluster)

. 
. generate KZ9b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum]
> *factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty 
> + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[undec]*undec_c 

. 
. generate Pkz9b=1/(1+exp(-KZ9b))

. 
. label var Pkz9b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz9b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predict
> ion) title(Modified Model Includes Interaction w/ Decade) saving(C:\Temp\Fig2g , replace)

. 
. 
. * Interaction with Treaty
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  untre, cluster(cluster)

. 
. generate KZ10b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum
> ]*factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty
>  + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[untre]*untre_c 

. 
. generate Pkz10b=1/(1+exp(-KZ10b))

. 
. label var Pkz10b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz10b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predic
> tion) title(Modified Model Includes Interaction w/ Treaty) saving(C:\Temp\Fig2h , replace)

. 
. 
. * Interaction with Trnsfcap
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  untra, cluster(cluster)

. 
. generate KZ11b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum
> ]*factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty
>  + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[untra]*untra_c 

. 
. generate Pkz11b=1/(1+exp(-KZ11b))

. 
. label var Pkz11b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz11b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predic
> tion) title(Modified Model Includes Interaction w/ Trnsfcap) saving(C:\Temp\Fig2i , replace)

. 
. 
. * Interaction with FACTNUM2
. 
. quietly logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap unintrvn treaty develop exp decade
>  unfac2, cluster(cluster)

. 
. generate KZ12b = _b[_cons] + _b[wartype]*wartype + _b[logcost]*logcost  + _b[wardur]*wardur + _b[factnum
> ]*factnum + _b[factnum2]*factnum2 + _b[trnsfcap]*trnsfcap + _b[unintrvn]*(1-unintrvn) +_b[treaty]*treaty
>  + _b[develop]*develop + _b[exp]*exp + _b[decade]*decade +_b[unfac2]*unfac2_c 

. 
. generate Pkz12b=1/(1+exp(-KZ12b))

. 
. label var Pkz12b "Modified Model Counterfactual Prediction"

. 
. set text 150

. 
. graph7 pds1 Pkz12b, l1(Original Model Counterfactual Prediction) b2(Modified Model Counterfactual Predic
> tion) title(Modified Model Includes Interaction w/ Factnum2) saving(C:\Temp\Fig2j , replace)

. 
. set text 100

. 
. graph7 using C:\Temp\Fig2a C:\Temp\Fig2b C:\Temp\Fig2c C:\Temp\Fig2d C:\Temp\Fig2e C:\Temp\Fig2f C:\Temp
> \Fig2g C:\Temp\Fig2h C:\Temp\Fig2i, margin(15) title("Counterfactual Predictions of PB Effects of UN Int
> ervention") saving(C:\Temp\Fig2Combo , replace)

. 
. 
. 
. ** Appendix 1 of the supplement includes simple numerical examples on the relationship between the odds 
> ratio and probability estimates in logit models
. ** This is related to our discussion in the ISQ paper of the fact that logit models are inherently inter
> active.
. 
. ** Appendix 2 of the supplement includes some 
. ** These figures are similar to the ones we presented in DS2000 (figures 2 and 3)
. ** Information on how to replicate the conditional effects plots in DS2000 (figures 2 and 3) was posted 
> online at the time of publication (2000)
. ** The do-files can be downloaded from this website:
. ** http://www.worldbank.org/research/conflict/papers/peacebuilding/Graphs_log.pdf
. 
. ** Supplement Figure A1: Effect of War Duration with and without a PKO and Treaty
. 
. logit pbs2s3 wartype logcost wardur factnum factnum2 trnsfcap un2cint treaty develop exp decade, robust 
> cluster(cluster)

Iteration 0:   log pseudo-likelihood = -79.172582
Iteration 1:   log pseudo-likelihood = -50.763958
Iteration 2:   log pseudo-likelihood = -46.061907
Iteration 3:   log pseudo-likelihood = -45.267339
Iteration 4:   log pseudo-likelihood = -45.236614
Iteration 5:   log pseudo-likelihood = -45.236556

Logit estimates                                   Number of obs   =        122
                                                  Wald chi2(11)   =      30.42
                                                  Prob > chi2     =     0.0014
Log pseudo-likelihood = -45.236556                Pseudo R2       =     0.4286

                          (standard errors adjusted for clustering on cluster)
------------------------------------------------------------------------------
             |               Robust
      pbs2s3 |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     wartype |  -2.083053   .6619669    -3.15   0.002    -3.380484   -.7856219
     logcost |  -.5345676   .1478955    -3.61   0.000    -.8244375   -.2446977
      wardur |   .0091107   .0055748     1.63   0.102    -.0018157    .0200371
     factnum |  -1.764165   .6858743    -2.57   0.010    -3.108454    -.419876
    factnum2 |   .1143719   .0565961     2.02   0.043     .0034455    .2252983
    trnsfcap |   .0038062   .0015776     2.41   0.016     .0007142    .0068983
     un2cint |   .7068036    .271515     2.60   0.009      .174644    1.238963
      treaty |   2.081591   .7632858     2.73   0.006     .5855782    3.577603
     develop |   .0005771   .0003616     1.60   0.111    -.0001317    .0012858
         exp |  -5.337883    2.98481    -1.79   0.074      -11.188    .5122377
      decade |  -.3490538   .1748891    -2.00   0.046      -.69183   -.0062775
       _cons |   10.31805   2.425895     4.25   0.000     5.563379    15.07271
------------------------------------------------------------------------------

. 
. ** Prob(PBS) given variable WARDUR and no UNops VS. multi PKO w/ treaty
. 
. generate L21 = _b[_cons] + _b[wartype]*1 + _b[logcost]*12.59473  + _b[wardur]*wardur +  _b[factnum]*3 + 
> _b[factnum2]*9 + _b[trnsfcap]* 14.31405 + _b[un2cint]*0    + _b[treaty]*0 + _b[develop]*221.4169 + _b[ex
> p]*.11804 + _b[decade]*4

. generate Phat21=1/(1+exp(-L21))

. label var Phat21 "No UN/Treaty"

. generate L22 = _b[_cons] + _b[wartype]*1 + _b[logcost]*12.59473  + _b[wardur]*wardur +  _b[factnum]*3 + 
> _b[factnum2]*9 + _b[trnsfcap]* 14.31405 + _b[un2cint]*3    + _b[treaty]*1 + _b[develop]*221.4169 + _b[ex
> p]*.11804 + _b[decade]*4

. generate Phat22=1/(1+exp(-L22))

. label var Phat22 "PKO & Treaty"

. graph7 Phat21 Phat22 wardur if wardur<350, connect(ss) ylabel xlabel l1(Prob(PB success)) b2(War Duratio
> n in Months) saving(C:\Temp\FigA1 , replace) 
(note: file C:\Temp\FigA1.gph not found)

. 
. ** Supplement Figure A2: Effect of War Duration in a Difficult Case w/ & w/o a PKO and Treaty
. 
. ** set deaths and displacements and number of factions at 75th pctl 
. ** so, this is equivalent to a "difficult case" in DS2000
. ** though we keep local capacities at their median levels given that there is no significant difference 
> with respect to war duration
. ** wartype set at "ethnic" for both cases, given that there is no significant difference with respect to
>  war duration
. 
. generate L23 =  _b[_cons] + _b[wartype]*1 + _b[logcost]*13.83109  + _b[wardur]*wardur +  _b[factnum]*4 +
>  _b[factnum2]*16 + _b[trnsfcap]* 1.963351 + _b[un2cint]* 0  + _b[treaty]*0 + _b[develop]*221.4169 + _b[e
> xp]* .1414055 + _b[decade]*4

. generate Phat23=1/(1+exp(-L23))

. label var Phat23 "No PKO/Treaty"

. generate L24 =  _b[_cons] + _b[wartype]*1 + _b[logcost]*13.83109  + _b[wardur]*wardur +  _b[factnum]*4 +
>  _b[factnum2]*16 + _b[trnsfcap]* 1.963351 + _b[un2cint]* 3 + _b[treaty]*1 + _b[develop]*221.4169 + _b[ex
> p]* .1414055 + _b[decade]*4

. generate Phat24=1/(1+exp(-L24))

. label var Phat24 "PKO & Treaty"

. graph7 Phat23 Phat24 wardur if wardur<350, connect(ss) ylabel xlabel l1(Prob(PB success)) b2(War Duratio
> n in Months) saving(C:\Temp\FigA2 , replace)
(note: file C:\Temp\FigA2.gph not found)

. 
. 
. 
. 
. 
. 
. 
. 
end of do-file

. log close
       log:  C:\Temp\DS2000replicationlog.log
  log type:  text
 closed on:  22 Apr 2006, 11:11:17
----------------------------------------------------------------------------------------------------------
